Momentum return: is it a compensation for risk?
Munira, S. (2009). Momentum return: is it a compensation for risk?. (Unpublished Doctoral thesis, City, University of London)
Abstract
This thesis examines if momentum returns are compensation for risk. Using a sample period from 1926 through 2006 for all stocks listed in the NYSE, AMEX and NASDAQ we provide a comprehensive analysis of momentum returns both at the portfolio and at the individual stock level, by using firm level and macro level risk factors and by employing contemporaneous and lagged values of risk factors. The study employs an alternative momentum strategy, measures the relative contribution of risks factor that generates momentum returns and establishes a link between momentum returns, uncertainty and credit ratings.
We report raw momentum returns of 0.8 percent per month (9.6 percent per annum) when returns are measured using the conventional methodology at the portfolio level. Momentum returns are predominantly high and earn more than 1 percent per month during the post-1950s compared to its counterpart in the pre-1950s. The study reports that when measured at the portfolio level momentum returns cannot be explained by risk factors. We document momentum returns of up to 0.01 percent per month (0.12 percent per annum) after Fama-French three factors, Carhart four factors and macroeconomic risk factors are priced for. The results are robust when the lagged values of these risk factors are employed. We further document momentum returns of 0.16 percent per month (1.92 percent per annum) when transaction cost is taken into account.
When measured at the individual stock level momentum returns cannot be explained by Fama-French three factors and contemporaneous values of macroeconomic risk factors. Unexplained returns are observed up to 0.45 percent per month (5.4 percent per annum) when Fama-French three factors are used. Unexplained returns up to 0.15 percent per month (1.8 percent per annum) are observed when contemporaneous values of macroeconomic risk factors are used. However, when the lagged values of macroeconomic risk factors are used, momentum returns disappear.
We decompose momentum returns to measure the relative contribution of the risk factors and the unexplained portion of momentum returns. At the portfolio level, decomposition shows that less than 10 percent of the contribution is from Fama-French three factors and less than 20 percent of the contribution is from macroeconomic risk factors. Unexplained portion contributes the remaining 90 percent and 80 percent, respectively. At the individual stock level, decomposition shows that contribution of both Fama-French three factors and macroeconomic risk factors increases up to 47 percent and 59 percent, respectively; unexplained portion contributes the remaining 63 percent and 41 percent, respectively. When lagged values are used the contribution of risk factors increases up to 68 percent.
Finally, we consider uncertainty at the firm level and the macro-economic risk level by measuring momentum returns of credit rated stocks. We observe momentum returns of 1.22 percent per month (14.64 percent per annum) in credit rated stocks. Among the credit rated stocks momentum returns are mainly earned by speculative grade stocks and during contractions. Momentum returns of about 2 percent per month (23 percent per annum) are observed in speculative grade stock and they are more pronounced of up to 4.99 percent per month (59.88 percent per annum) during contractions. However, momentum returns of speculative grade stocks disappear when controlled for macroeconomic risk factors.
We show that momentum is quite persistent when measured at the portfolio level by using the conventional approach and at the individual stock level when using an alternative approach. Momentum returns cannot be explained by Fama-French three factors and contemporaneous values of macroeconomic risk factors. However, only at the individual stock level, lagged values of macroeconomic risk can explain momentum returns. When we decompose momentum returns into explained and unexplained components, we provide support to the above findings that the contribution of macroeconomic risk factors is the highest when measured at the individual stock level. Momentum are reactions of the investors’ to high uncertainty, when uncertainty is measured at the firm level or at the macro level by measuring returns of credit rated stocks. Momentum returns are investors’ reaction due to increased business risk of stocks or due to increased macroeconomic risk during downturns.
It can be concluded that at the portfolio level momentum returns remain when risk factors are price for and at the individual stock level momentum returns diminishes though they do not disappear entirely when risk factors are controlled for. When momentum returns are decomposed at the portfolio level unexplained risk factors contributes the most and at the individual stock level contribution of risk factors increases among which the contribution of macroeconomic risk factors increases the most. Momentum returns could be a compensation for uncertainty at the firm level as it concentrates mostly on Speculative Grade rated stocks with more pronounced effect during contraction. Momentum returns disappear when macroeconomic risk factors are priced.
Publication Type: | Thesis (Doctoral) |
---|---|
Subjects: | H Social Sciences > HG Finance |
Departments: | Bayes Business School > Finance |
Download (761kB) | Preview
Export
Downloads
Downloads per month over past year