Tests performed did not find significant differences in mean values of profits, assets and
revenues between samples. However, we found significant differences in median values
of assets with p<0.01 and revenues with p<0.1, that is consistent with the application of
different valuation methods. Results suggest that the use of fair value for biological
assets significantly changed the value of farm assets and revenues, with respect to
historical cost valuation,but did not significantly affect the amount of profits.
The absence of significant differences in standard deviation of profits, assets and
liabilities does neither provide empirical evidence, for the agricultural sector, to the
commonly accepted hypothesis (e.g. Plantin and Sapra, forthcoming; Dowling, 2001;
Pentinen et al., 2004) of greater volatility with fair valuation, nor to Bleck and Liu’s
(2007) hypothesis of greater volatility with historic cost. No significant differences in
coefficients of variation of these variables confirm this result.
The fact that mean and median values of return on asset are not significantly different
between groups neither confirm, in the specific circumstances of agricultural sector,
Liang and Weng’s (2007) hypothesis of less efficient decisions under fair valuation, nor
Black and Liu’s (2007) argument that under historic cost bad investment projects would
be pooled with good projects and prevented from liquidation. In a similar way, no
significant differences in standard deviation of return on assets confirms the absence of
greater volatility under fair value and suggests that there is no significant transfers of
gains and loses between periods.
No significant differences in standard deviation of profits to standard deviation of cash
flows between groups suggests that fair value did not provide greater discretionarily to
manipulate earnings, as is usually assumed (e.g. Watts, 2003; Liang and Wen, 2007).
Regressions performed for equations (1), (2) and (3) are displayed in table 2. Column
(A) displays estimations when the independent variable is standard deviation of profits,
Values of variance inflation factors, condition indexes and variance proportion of
variables suggest that col linearity and multicollinearity do not likely disturb regression
estimations. The model presents a significant goodness-of-fit and explains about 98% of
the total variability. The insignificant sign for the coefficient of the dummy variable
suggests no influence of the valuation method in profit volatility. The control variable
presents the expected significant positive sign.
The Durbin-Watson statistics for regression estimations of equations (2) and (3)
determine the typical autocorrelation pattern for independent variables in panel data.
Given that we have unbalanced panel data, we performed panel data regressions with
7
random effects. Results displayed in columns (B) and (C) of table 2 reinforce the
absence of influence of fair valuation in volatility of profits. Both models present a
significant goodness-of-fit and the expected significant positive sign for the control
variables. In none of the columns the dummy variable for valuation method presents a
significant sign.
Therefore, estimations from table 2 reinforce our finding that fair valuation of biological
assets is not associated with higher volatility.
การแปล กรุณารอสักครู่..
