As seen in Table 1, we find some studies support the positive impact
of stock-based compensation. For example, Larcker et al. (2007) show
that both cash-based and stock-based compensation have significantly
positive effects on a firm’s future performance, and this result is consistent
within market and accounting performance measures. In addition,
Aboody et al. (2010) test the incentive effect related to re-pricing
underwater options, and find that firm performance, as measured by
operating income or cash flows, is substantially improved. This means
that re-pricing restores the incentive effect in the stock option grants,
which is similar to the incentive effect that occurs with new grants of
stock-based compensation.
On the other hand, some researchers show the negative impact of
stock-based compensation on firm future performance. Core et al.
(1999) find that firms with greater excess compensation perform poorly
relative to those with lower or negative excess compensation. Firms
with weaker governance structures are more likely to have excess compensation.
Therefore, excess compensation resulting from agency problems
has a negative influence on both operating and market
performance. Brick et al. (2006) and Cheng and Farber (2008) provide
similar results with slightly different incentive and performance measures.
Mixed findings on the association between CEO equity compensation
and firm performance have generated considerable interest, with
two especially interesting directions of research. The first investigates
the conditional relation between CEO equity compensation and firm
performance. The second direction examines whether the association
between CEO equity compensation and performance is robust to the
use of the econometric model. In an earlier research survey, Ehrenberg
and Milkovich (1987) mention that some proxies for firm’s performance,
such as firm size or others, are considered in terms of their
effects on a firm’s compensation strategies. In later empirical works,
researchers use simultaneous equations methods to model the endogenous
relationship between firm performance and equity compensation.4
To address the conditional relation between CEO equity compensation
and firm performance, researchers try to use different segmentations
of firms to re-test the incentive effects. Following Murphy (2003),
and Ittner et al. (2003), Kim (2010) finds that new economy, young, or
volatile firms provide a higher incentive from stock-based compensation.
Kim also shows evidence that the major driver for the variation
of the incentive effect is option grants. Hence, to be consistent with theexisting literature, we focus on the stock-based components in the
analysis of the incentive effect in executive compensation.
In contrast to segmentation by firm characteristics, Matolcsy (2000)
analyzes the pay-for-performance sensitivity for cash compensation
using a sample of 100 randomly selected Australian Statex companies.
Within different stages of the economic cycle, the sensitivity of cash
compensation to changes in net profit is asymmetric. Similarly, Frye
(2004) tests the relationship between stock-based compensation and
firm performance by using firm data in different time periods, and
finds a positive impact of stock-based compensation on firm ROA in
the early 1990s, but a negative impact in the late 1990s.
Aggarwal and Samwick (2006) show that Tobin’s q increases with
incentives across different levels of incentives. They use 5 and 25 per
cent as the critical values to separate firms into different segments with
respect to the incentives offered to top managers. Even though the
authors find consistent results for all levels of incentives, it is still possible
that different levels of incentives or firm performance measures
may be an important reason for the different impact of stock-based
compensation on firm performance that is found in the literature.
Following a similar idea, Canarella and Nourayi (2008) utilize the
sign of performance measures and apply the GMM estimation to test
the asymmetric incentive effect in total and cash compensation. They
find a non-linear and asymmetric relationship between executive compensation
and firm performance, and that the non-linearity and asymmetry
depends on the measures of firm performance. Convexity
characterizes the asymmetry of the relationship between executive compensation
and market returns, while concavity distinguishes the asymmetry
of the relationship between executive compensation and
accounting returns.
As shown above, many researchers find that there is an asymmetric
impact of executive compensation on firm performance, and provide
various possible explanations for this. Essentially, they use different
data segmentation methods to identify the factors that could influence
the incentive effect in stock-based compensation. However, in general,
these segmentations are arbitrary and exogenous, and normally they
would change the original distribution of the full sample, which might
invalidate the statistical tests. The QR method is one way to alleviate
these problems, and thus we apply it to test the relationship between
stock-based compensation and firm performance.