Read the article, Combine those observations with the Figure2,3 and Table 4,5 ,
ID: 374545 • Letter: R
Question
Read the article, Combine those observations with the Figure2,3 and Table 4,5 , briefly summarize it
Observation 2Poor matching increases the volatility of earnings
Observation 3: Poor matching decreases the persistence of earnings.
Observation 4: The effects of poor matching are alleviated over longer-time horizons.
Next, we turn to documenting the effect of poor matching on earnings volatility. From Observation 2 and the conjecture that matching has become worse over time, we expect to see that earnings volatility has increased over time. To reflect the fact that earnings volatility arises from the volatility of the underlying fundamentals in addition to poor matching, we also include accompanying results for revenues and expenses. Specifically, since earnings is equal to revenues minus expenses, the volatility of earnings is equal to the volatility of revenues plus volatility of expenses, adjusted for the correlation between revenues and expenses, so we include the data for all these variables. The value of this analysis is that (as discussed above) the correlation between revenues and expenses reflects the quality of matching, while the volatilities of revenues and expenses can be thought of as capturing the volatility of the underlying business fundamentals, and thus one can assess whether the increase in earnings volatility is due to increase in the volatility of fundamentals.
Table 4, Panel A contains the results for the one-year specification, providing mean values for each year in the sample and t-tests for differences in means between the first and second half of the sample period. These results are also graphically depicted in Figure 2, Panel A, where all values are normalized to a beginning value of 100 and smoothed over five years to enhance comparability. An inspection of Figure 2 and Table 4 reveals that earnings volatility has substantially increased over the last four decades, and that this increase is highly statistically significant. The inspection also reveals that the increase in earnings volatility cannot be attributed to increase in the volatility of fundamentals because there is no increase and an actual slight decrease in the volatilities of revenues and expenses. In contrast, there is a material and highly statistically significant decrease in the correlation of revenues and expenses over time. The joint consideration of these two results implies that the trend of increasing earnings volatility over time is entirely due to poor matching.
Note that the results for increasing volatility of earnings are not new. Givoly and Hayn 2000 find a great temporal increase in earnings volatility, interpreting it as due to the increase in reporting conservatism over time. While our interpretation is based on temporal deterioration of the matching process, it is useful to point out that these two perspectives share much of the same underlying process. In essence, conservatism can be viewed as a form of “poor matching,” where the expenses precede the associated revenues. Thus, these two interpretations are often just two different lenses through which to view the same effects, and whether one adopts one versus the other partly depends on the goals and the particular needs of the user. We adopt the matching lens because it has not been explored before and because it allows us to offer insights unavailable from the existing conservatism perspective. For example, the matching perspective allows a clear explanation for the substantial increase in the correlation between current revenues and future expenses, which is the opposite of what increasing conservatism would predict.
Table 4, Panel B presents the corresponding volatility results for the two-year specifications of all variables. From Observation 4 and the intuition that the effects of poor matching resolve over time, we expect that the temporal patterns of increased volatility of earnings and decreased contemporaneous correlation between revenues and expenses are less-pronounced for the two-year specifications. This expectation is partially borne out in the data. The decline in the correlation between revenues and expenses in Panel B is only half of that in Panel A, and this difference is significant at the 0.001 level based on bootstrap tests as in Noreen 1989. 5 Thus, there is clear evidence that the decline in the quality of matching is less pronounced for longer-horizon definitions. The increase in volatility of earnings is also smaller for Panel B as compared to that in Panel A but in this case the difference is minimal 0.006 versus 0.007 and not statistically significant bootstrap p-value=0.240. The most likely reason for the minimal difference in volatility is that our longer horizon is only two years. Thus, the included two-year results should be more properly considered as the lower limit on how the consideration of longer-horizon variables attenuates the effects of poor matching.
The results for persistence of earnings and autocorrelation in earnings changes are presented in Figure 3 and Table 5, including one-year and two-year specifications. Based on Observation 3 and its Corollary, we expect that earnings persistence has declined over time and that the autocorrelation in earnings changes has become more negative. An informal inspection of the smoothed one-year results in Figure 3 , Panel reveals a clear support for this conjecture, with steady downward drift in both variables over time. The numerical results and tests in Panel A in Table 5 also confirm this pattern. The mean earnings persistence over the first half of the sample period is 0.855, while it is only 0.705 over the second half of the sample, for a difference of 0.15, which is highly statistically significant. The mean autocorrelation in earnings changes is only 0.019 in the first half but becomes 0.234 in the second half, for a difference of 0.215, which is highly statistically significant.
We aim to provide some additional feel for the economic importance of these results by using a regression of the yearly earnings persistence and autocorrelation variables on an ordinal time variable, where the calendar year is replaced by its time rank e.g., the first year 1967 is 0 and the last year 2003 is 36. The regression results in Panel B of Table 5 confirm that earnings persistence is declining over time, with a negative and highly significant coefficient on the time variable. More relevant in terms of economic importance, the regression reveals an adjusted R 2 of 0.54, which suggests that this downward trend is the defining feature of the evolution of earnings persistence over time, accounting for more than half of its time-series variation. The fitted value for earnings persistence for the first year of the sample 1967 is 0.91 and the corresponding number for the last year is 0.65, which reveals an economically substantial decline in earnings persistence over the last four decades. We repeat the same regression analysis for the autocorrelation in earnings changes. The R 2 is 0.39, which suggests that more than a third of the total temporal variation in the autocorrelation variable is due to its secular decline over time. The fitted value for the first year is 0.053 and the one for the last year is 0.299. Thus, while the beginning value is close to zero, by the end of the sample, earnings exhibits economically strong reversal in changes, where nearly a third of the typical change is immediately reversed in the next period.
The two-year results in Figure 3 and Table 5 provide further evidence whether the effects of poor matching are alleviated for longer-horizon variables. An informal inspection and comparison of Figure 3, Panels A and B, indicates that this is likely the case because the pronounced downward drifts in Figure 3, Panel A are attenuated in Figure 3, Panel B. Turning to the numbers in Panel C of Table 5, we find further support for this conjecture. The difference in means between the first half and the second half of the sample is 0.111 for the two-year specification as compared to 0.150 in Panel A, and this attenuation is marginally significant bootstrap p-value = 0.09. The results are much more emphatic for the autocorrelation in earnings changes, where the time-series difference in means in Panel C is only 0.058 as compared to 0.215 in Panel A, an attenuation which is both statistically and economically significant with a bootstrap p-value of 0.001. As already discussed, this pattern of attenuation in two-year results should be probably thought of as a lower limit on how the effects of poor matching are attenuated in longer-horizon specifications of variables.
TABLE 4 Volatility of Earnings and Its Components over Time Panel A: Volatility over Time for the One-Year Sample Year 1967 Corr (Revenues, Expenses) Vol (Earnings)Vol (Revenues)Vol (Expenses) 0.099 0.092 0.972 1969 1970 1971 1972 0.978 0.973 0.966 0.968 0.969 0.973 0.979 883 0.105 0.097 957 957 955 957 957 1974 1975 1976 1977 1978 1979 1980 1981 1982 1983 1984 1985 1986 1987 1988 1989 1990 0.082 0.098 0.095 0.094 0.088 0.089 0.094 0.090 0.089 0.089 0.092 0.971 0.972 0.978 0.976 0.974 0.970 0.970 0.968 959 958 0.083 0.087 0.094 0.097 0.095 0.090 953 957 0.097 958 955 955 0.951 0.943 0.933 0.924 0.909 0.907 0.905 0.020 0.020 0.020 0.021 0.020 0.020 0.020 0.020 0.020 0.021 0.021 0.023 0.026 0.029 0.099 0.089 0.083 0.075 0.069 0.066 0.070 0.073 0.081 0.089 0.093 0.097 958 0.089 0.080 0.073 0.069 0.073 1992 1993 959 1995 1996 1997 1998 957 957 960 0.892 0.895 0.901 0.904 0.920 2000 2001 2002 0.098 0.916 (continued on next page)Explanation / Answer
Option B: Single-use plan
A programme is a single use plan to carry out a special project within an organization. Once the project is over and new project is assigned than new plan should be implemented. So a programme is a single use plan.
Related Questions
drjack9650@gmail.com
Navigate
Integrity-first tutoring: explanations and feedback only — we do not complete graded work. Learn more.