International Review of Business Research Papers

Vol. 9. No. 6. November 2013 Issue. Pp. 102 – 110

Dynamic Capital Structure Trade-off Theory: Evidence from

Malaysia
Islam Abdeljawad*, Fauzias Mat-Nor**, Izani Ibrahim***
and Ruzita Abdul-Rahim****
Dynamic trade-off theory proposes that firms may deviate from their target capital
structure but they will exhibit an adjustment behavior towards that target.
Estimating the speed of adjustment (SOA) is an investigation for the joint
hypotheses that the target actually exists and that firms adjust toward their target.
However, using a single estimated SOA that fits all firms is misleading and cannot
be used as evidence for or against the dynamic trade-off theory if the real
dynamic behavior is largely heterogeneous. This study finds evidence for this
heterogeneity in the SOA for Malaysian firms using system GMM approach. The
study finds that firms that are far from the target exhibit faster adjustments than
firms close to the target, and firms that are overleveraged exhibit faster
adjustment than underleveraged firms. These results are consistent with the
dynamic trade-off theory. However, the finding of this study cannot reject other
interpretations of capital structure i.e., pecking order or timing theory, since the
SOA for some of the firms is very slow and hence, cannot be the first order
determinant of capital structure.

JEL Codes: G32, G33, and O16

1. Introduction
Trade-off theory of capital structure (Baxter 1967; Kraus & Litzenberger 1973)
suggests that firms choose their capital structure by balancing the advantages of
borrowing, mainly tax savings, with the costs associated with borrowing including
bankruptcy costs. This trade-off implies the existence of a target leverage that
maximizes the value of the firm. The existence of a target, which is at the heart of the
theory, requires that any deviation from that target leverage should be adjusted.

The dynamic version of the trade-off theory explicitly accounts for the adjustment
behavior of the leverage ratio where adjustments take place when the cost of deviation
from the target exceeds the cost of adjustment towards that target (Fischer et al.
1989). One advantage of the dynamic feature is that since the adjustment towards the
target is a characteristic of trade-off theory, it can be used to validate the trade-off
theory against other theories of capital structure that do not presume the existence of
target leverage, i.e. pecking order theory (Myers & Majluf 1984) and timing theory
(Baker & Wurgler2002).

* Assist. Prof. Dr. Islam Abdeljawad, Faculty of Economics and Administrative Sciences, An-Najah National
University, Nablus, Palestine. Email: [email protected] Tel: +9729-2345113, Fax: +9729-2345982
** Prof. Dr. Fauzias Mat-Nor, Graduate School of Business, National University of Malaysia, Selangor,
Malaysia. Email: [email protected] Tel: +603-89213792, Fax: +603-89213161
*** Prof. Dr. Izani Ibrahim, Graduate School of Business, National University of Malaysia, Selangor, Malaysia.
Email: [email protected] Tel: +603-89213004, Fax: +603-89213161
**** Assoc. Prof. Dr. Ruzita Abdul-Rahim, Faculty of Economics and Management, National University of
Malaysia, Selangor, Malaysia. Email: [email protected] Tel: +603-89215764, Fax: +603-89213163
 Abdeljawad, Mat-Nor, Ibrahim & Abdul-Rahim
In both pecking order and timing theories, the leverage dynamism is driven by factors
that are not related to the target leverage i.e., adverse selection costs and mispricing
of the firm’s equity, respectively. As argued by Fama and French (2002), the existence
of the target and the adjustment toward that target are the most convincing evidence in
favor of trade-off theory.

The two features of trade-off theory, namely the existence of the target and the
adjustment toward that target, can be jointly tested by estimating the speed of
adjustment (SOA) to the target. The SOA is the percentage of the deviation from the
target that the firm removes each period. In estimating the SOA, the implicit
assumption is that the SOA is homogenous across firms (Fama & French 2002;
Flannery & Rangan 2006; Ozkan 2001). This assumption is however inconsistent with
the argument of the dynamic trade-off theory which posits that different costs of
deviation and different costs of adjustments should result in different estimations for
the SOA. Few recent papers indicate that no single estimated SOA can fit all firms. A
more reasonable approach is to acknowledge multiple SOAs to understand the true
dynamic behavior of the firms (Flannery & Hankins 2007; Lemmon et al. 2008). The
heterogeneity of the SOA does not only give support to the trade-off theory; it may also
explain possible conditions under which other theories may prevail. In cases where the
adjustment speed is slow, the adjustment behavior may not be the priority such that
other considerations, for instance, the mispricing of equity or the adverse selection
costs may drive the capital structure decisions. Nonetheless, if the SOA is high, it is
likely that the adjustment to the target will dominate other considerations in which the
trade-off theory is likely to appear as the first order determinant.

This paper builds on this new line of arguments and aims to add evidence for the
heterogeneity of the SOA from Malaysian data. This paper conditions the deviation
from the target using two different criteria specifically, whether the firm is over or
under-leveraged and whether the firm is close or far from the target. This study
addresses the estimation problems highlighted in previous literature by using the
system GMM approach of Blundell and Bond (1998). This estimator is recent to the
Malaysian capital structure research and the results may improve the understanding of
the way Malaysian firms make financing decisions.

In the remaining of this paper, the review of the literature is presented in Section 2.
Section 3 discusses the methodology and estimation models, followed by Section 4
which reports and discusses the results and Section 5 which concludes.

2. Literature Review
Theories of capital structure can be grouped into two categories; the first includes
trade-off theory (Baxter 1967; Kraus & Litzenberger 1973), agency theory (Jensen &
Meckling 1976) and free cash flow theory (Jensen 1986) which recognize the
existence of an optimal level of debt (target leverage). The second category includes
pecking order theory (Myers & Majluf 1984) and equity market timing theory (Baker &
Wurgler 2002) that do not assume any optimal level of debt. Theories in both
categories can be modeled into static or dynamic framework. Adjustment behavior
towards the target is associated with the dynamic versions of theories from the first
category while the dynamic versions of theories from the second category do not
include target leverage that the firm is adjusting to. Instead, the dynamism of theories

103
 Abdeljawad, Mat-Nor, Ibrahim & Abdul-Rahim
from the second category relates to non-target factors like the cost of adverse
selection (pecking order theory) or the mispricing of the common stock (timing theory).

The main difference between the dynamism of theories in both categories is that the
dynamism implied by the first category theories gradually moves the observed
leverage toward the target leverage. Meanwhile, the dynamism of the second category
theories may move the observed leverage to any direction depending on the factor that
drives the leverage ratio. The present study intends to investigate the dynamic
adjustment towards the target. Hence, it is more related to theories of the first
category. Introducing the main characteristics of static versions of these theories may
facilitate the discussion about the dynamic adjustment toward the target later.
Interestingly, these theories can be seen as complementary to each other, not
substitutes (Myers 2003). Therefore, the costs and benefits in the real world can be the
sum of the costs and benefits highlighted in all theories.

The static version of capital structure theories that include target leverage emphasize
the idea that firms trade-off between costs and benefits of debt, and this trade-off can
explain the cross-section variation in leverage ratio across firms. These theories
implicitly assume the existence of target leverage but believe that all firms are already
at their targets. However, what determine the equilibrium point in these theories differs
from one theory to the other. In the tax-based trade-off theories, the benefits of interest
tax shield are offset by the costs from additional borrowing, particularly the bankruptcy
costs. Baxter (1967) and Kraus and Litzenberger (1973) state that a taxable
corporation should increase its debt level until the marginal value of tax shield is offset
by the present value of possible financial distress costs.

Dynamic trade-off theories explicitly emphasize the idea that firms have a target that
maximizes its value and deviations from that target are costly. Hence, deviations will
be gradually removed over time. Extant literature generally supports the existence of
long-run target leverage and agrees to the notion that a typical firm converges to that
target gradually at a certain SOA but the magnitude of this SOA is not a settled issue
(Frank & Goyal 2007). A very high SOA of about 80 percent yearly is found by de
Miguel and Pindado (2001). Ozkan (2001) also finds a relatively fast speed of more
than 50 per annum while Flannery and Rangan (2006) document a rapid but more
reasonable SOA of 35 percent yearly which they interpret as evidence in favor of the
trade-off theory. In contrast, Fama and French (2002) find a slow SOA that ranges
between 7 percent minimum for dividend payers and 18 percent maximum for dividend
non-payers while Huang and Ritter (2009) find a SOA of 17 percent annually. Despite
the slow SOA, Fama and French (2002) interpret their result as being consistent with
trade-off model but conclude that the result cannot be used to reject the pecking order
model. A slow SOA indicates that trade-off factors may be only a secondary
consideration in the capital structure decisions. Malaysian literature finds a relatively
active adjustment behavior that is usually interpreted in favor of the trade-off theory
(Hussain 2005). However, what is the reason for the high heterogeneity in the
estimated SOA? Some researchers point to fundamental reasons (Flannery & Hankins
2007) while others attribute the differences to the methodology used (Hovakimian &
Guangzhong 2011; Iliev & Welch 2010).

One possible reason for the heterogeneity of the SOA is the variation across countries
such as in terms of the country’s economic environment, institutions, tax systems and
governance practices on the capital structure decisions (Antoniou et al. 2008).

104
 Abdeljawad, Mat-Nor, Ibrahim & Abdul-Rahim
Nonetheless, the SOA is still likely to be heterogeneous even across firms within the
same country and in studies that use the same estimation method. Flannery and
Hankins (2007) suggest a theory of dynamic capital structure where capital structure
decisions reflect the level of optimal leverage as well as the cost of deviation from that
leverage and the costs of adjusting toward that optimal level of leverage. The SOA
depends on the costs and benefits of rebalancing. Since these costs vary across firms,
so do the SOAs. Elsas and Florysiak (2011) find that the SOA is higher for firms with
high default risk, high expected bankruptcy costs and higher opportunity costs of
deviating from the target. Drobetz and Wanzenried (2006) find that faster growing firms
and firms that are more deviated from the target adjust more quickly. Byoun (2008)
finds that a single constant SOA cannot capture the dynamics of capital structure for
all firms as financial deficit and over- or underleveraged of the firm may affect the
SOA. Byoun (2008) finds that firms that are above the target debt with financial surplus
adjust at a rate of 33 percent and firms that are below target with financial deficit adjust
at 20 percent. On the other hand, firms with deficit and above target adjust at 2 percent
while those with surplus and below target adjust at 5 percent. Meanwhile, Dang et al.
(2012) find that firms with large financing imbalances, large investment or low earnings
volatility adjust faster to their target capital structure.

3. Methodology
The study includes all non-financial firms listed on Bursa Malaysia for which data are
available from Thomson Financial Worldscope Database during the period of 1992-
2009. Data prior to 1992 is often missing and hence, returning to earlier year is not
feasible. Financial firms are excluded since their capital structure reflects special
regulations. Effect of outliers is restricted by excluding firm-year observations where (i)
book value of assets is missing, (ii) book leverage ratio exceeds 1.0, and (iii) market-
to-book ratio exceeds 10.0 (Baker & Wurgler 2002; Hovakimian 2006). The resulting
unbalanced panel data of 434 firms provides 7978 firm-year observations.

The dependent variable is the firm’s leverage ratio which is the ratio of total debt to
total assets. This leverage reflects only the debt financing policy of the firm
(Hovakimian 2006). Following Fama and French (2002), this study uses the book
value of leverage because it better captures the active adjustment behavior since
market leverage also captures adjustment to market fluctuation.

Standard partial adjustment model is used to capture the dynamic adjustment toward
the target, that is;

( )
(1)

where  is the average SOA to the target each period for all firms used in the
estimation, is the target leverage whereas and are the current and
lagged 1 period leverage ratios, respectively. The model assumes that the firm has a
target leverage that minimizes the cost of capital. Once the firm is deviated from the
target, it should adjust once the cost of deviation is higher than the cost of adjustment.
Since a full adjustment occurs when =1 while =0 means no adjustment takes place,
the partial adjustment model proposes that actual adjustment in leverage should be
between 0 and 1.

105
 Abdeljawad, Mat-Nor, Ibrahim & Abdul-Rahim
The target leverage is unobservable and hence, is proxied by the fitted value
from a regression of observed leverage on a set of firms’ characteristics that determine
target capital structure (Fama & French 2002; Flannery & Rangan 2006). The firm’s
characteristics used in this study are the variables found to be the significant capital
structure determinants in the US and other developed countries (Rajan & Zingales
1995; Baker & Wurgler 2002), as well as in Malaysia (Booth et al. 2001). The
regression results later show that the variance inflation factors (VIF) of these variables
range from 1.017 to 1.034.

(2)

Where and are firm and time fixed effects following Flannery and Rangan (2006)
and Lemmon et al. (2008). Other variables are self-explanatory (refer to Table 1).
However, an estimation of the partial adjustment model can be done by substituting
the target leverage from Eq. (2) into the partial adjustment model in Eq. (1) and by
rearranging the terms the estimation can be done in a single step of Eq. (3).

( ) (3)

Where is the SOA, equals to 1.0 minus the coefficient of the lagged leverage. is
the set of determinant variables of the target from Eq.(2). All variables are used as
concurrent regressors except for the lagged leverage, to allow for more observations
to be used.

To estimate the SOA, this study employs the system generalized methods of moments
(GMM) following recent studies (Antoniou et al. 2008; Lemmon et al. 2008). Lemmon
et al. (2008) argue that system GMM is expected to produce large efficiency gains
over other approaches that use difference GMM or two stages least square methods.
This study also uses orthogonal deviation to remove the fixed effects (Arellano &
Bover 1995; Roodman 2006). Orthogonal deviations are preferable in cases where
many gaps exist in the unbalanced panel data, especially when sub-samples are
created. That is, after Eq. (3) is estimated for the full sample, it is re-run on sub-
samples of over or underleveraged firms to control for the direction of the deviation
from the target. Investigating the differences in capital structure determinants using
partial adjustment models by dividing the full sample into subgroups can be found in
Antoniou et al. (2008). Finally, Eq. (3) is re-estimated for another set of two sub-
samples of firms close to the target and firms far from the target. Deviation (DEV) from
the target is;

(4)

where is estimated using Eq. (2). Since leverage value is by definition bounded
by minimum 0 and maximum 1, any fitted value for the target leverage that is out of the
sample observations is replaced by its actual value to be consistent with the defined
boundaries. The sub-samples are then created using the median of the absolute value
of DEV as the cut-off point. Firms below the median (near-target firms) are separated
from those above the median (off-target firms).

106
 Abdeljawad, Mat-Nor, Ibrahim & Abdul-Rahim
4. Estimation Results
The descriptive statistics for the variables are presented in Table 1. Results from
estimating Eq. (3) for the full and sub-samples are reported in Table 2. For diagnostics,
both significance of AR(1) and AR(2) are reported and the insignificant AR(2) indicates
the absence of second order autocorrelation as required for a GMM estimation. The
validity of instruments is satisfied under the Hansen test. The Wald test indicates that
the regressors are jointly significant in explaining the dependent variable. The number
of instruments is also reported following the recommendation of Roodman (2006).

Table 1: Descriptive statistics

Variable Definition Mean Median Maximum Minimum Std. Dev.
Leverage Total Debt/Total Assets 0.251 0.233 0.997 0.000 0.196
Growth MV(Equity)/BV(Equity) 1.116 0.918 9.297 0.000 0.737
Tangibility Net PPE/Total Assets 0.403 0.394 0.999 0.000 0.222
Size Log(Sales) 19.104 19.026 23.969 9.250 1.545
Profit EBITD/Total Assets 0.067 0.075 11.096 -2.434 0.188
Notes: N is always equals to 7978 except for size where N equals to 7955. PPE is net plant, property
and equipment and EBITD is earnings before interest, taxes and depreciation.

The results for the full sample in Column (a) of Table 2 show that the lagged leverage
is the most important determinant of current leverage. Holding all other regressors
constant, about 87.4 percent change in the mean of current leverage is resulted from a
100 percent change in the lagged leverage. This is evidence for the high persistence
of the leverage variable. The lagged dependent variable is of special importance in the
partial adjustment model. If the firm follows an adjustment policy, the coefficient of this
variable must lie between 0 and 1.

The speed of adjustment (SOA, ) for the full sample suggests that only 12.7 percent
of the difference between desired and actual level of leverage is closed in one year.
The low SOA is consistent with results reported in developed markets (Baker &
Wurgler 2002; Fama & French 2002; Huang & Ritter 2009; Iliev & Welch 2010;
Lemmon et al. 2008). However, this result is different from Flannery and Rangan
(2006) who document a relatively high SOA of 34.4 percent. The difference can be
traced to the estimation approach used. Flannery and Rangan (2006) use a within-
estimator to remove the fixed effects, which tends to be severely biased in estimating
dynamic models (Lemmon et al., 2008). The slow SOA is also inconclusive in
supporting or rejecting the trade-off theory. The adjustment behavior is significant but it
is too small to be a first priority in determining the capital structure. Fama and French
(2002) find similar slow SOA ranging from 7-17 percent and they find it difficult to
interpret the results in favor of the trade-off theory despite the statistically reliable
coefficient.

Overleveraged firms face higher costs of deviation so that they are more pressured to
adjust. Meanwhile, the costs of deviation are lower for underleveraged firms where
adjusting is not a priority for these firms. To investigate this hypothesis, two
subsamples representing overleveraged and underleveraged firms are created. Eq. (3)
is re-estimated for each subsample and the results are reported in Columns (b) and (c)
of Table 2. Generally, firms are found to be much more sensitive when they are
overleveraged than when they are underleveraged. The SOA for overleveraged firms

107
 Abdeljawad, Mat-Nor, Ibrahim & Abdul-Rahim
is much higher (29.4%) than for underleveraged firms (13.1%). The impacts of control
variables are also much higher for overleveraged than for underleveraged firms.

Table 2: Estimation Results for the Short-Run Determinants of Leverage

Off-target Near-target
Whole Over-levered Under-levered
Leverage Leverage
LEV(-1) 0.8743 0.7061 0.8687 0.8253 0.9767
(36.91)*** (9.41)*** (21.08)*** (27.22)*** (9.41)***
Growth 0.0117 0.0201 0.0029 0.0164 0.0037
(3.59)*** (4.62)*** (1.09) (3.10)*** (0.67)
Profit -0.2156 -0.2606 -0.1259 -0.2370 -0.0679
(-8.61)*** (-8.04)*** (-3.93)*** (-5.83)*** (-0.96)
Size 0.0049 0.0045 0.0026 0.0051 -0.0004
(4.23)*** (2.35)** (2.14)** (2.08)** (-0.48)
Tangibility 0.0365 0.0477 0.0150 0.0428 0.0140
(5.04)*** (3.26)*** (2.35)** (3.14)*** (0.58)

SOA 12.7% 29.4% 13.1% 17.48% 2.3%

No. of Obs. 6931 2384 2644 2489 2496
No. of Inst. 156 156 154 156 152
Sig. of AR(1) 0.000 0.000 0.000 0.000 0.000
Sig. of AR(2) 0.989 0.679 0.942 0.502 0.937
Sig. of
0.668 0.352 0.454 0.752 0.684
Hansen test
Wald Chi2 (2579.8)*** (888.9)*** (1169.2)*** (1251.5)*** (72732.8)***
Notes: SOA is speed of adjustment. Constant coefficient and time dummies are included but not
reported. Standard errors are robust and corrected using Windmeijer’s (2005) finite sample
correction. The significance of Arellano and Bond’s (1991) test for AR(1) and AR(2) are reported. ***,
** and * indicate the coefficient is significant at 1%, 5% and 10% levels, respectively.

Next, the magnitude of the deviation is controlled for by creating two sub-samples;
firms close to the target and firms far from the target. Dynamic trade-off theory expects
that the higher the costs of deviation, the faster the adjustment speed. Columns (d)
and (e) of Table 2 present the results. The SOA for firms that are far from the target is
about 17.5 percent while it is only 2.3 percent for firms close to the target. This result is
consistent with the predictions of dynamic trade-off theory. Table 2 also reveals that all
the determinants of leverage ratio are only significant for the off-target leverage firms.
The large coefficient of the lagged leverage for firms close to the target indicates that
the leverage ratio is very persistence for these firms.

5. Conclusion
Using system GMM, this study reveals that Malaysian firms are adjusting their capital
structure to the target but at a slow rate of 12.7 percent. However, this study finds
evidence for heterogeneity in the targeting behavior; overleveraged firms adjust to the
target faster than underleveraged firms (29.4% versus 13.1%). This behavior is likely
to result from the asymmetry of the benefits of being at the target. Deviating from the
target on the upper side is likely to be more costly than deviating below the target
because bankruptcy costs and agency costs of debt will intensify quickly as the firm
deviates more above the target. Hence, overleveraged firms need to adjust faster to
reduce these costs. The dynamic behavior for firms far from the target and firms close
to the target is different which is expected as greater deviation from the target makes it
more critical for the firm to adjust. Firms far from the target are found to adjust faster

108
 Abdeljawad, Mat-Nor, Ibrahim & Abdul-Rahim
(17.5%) than firms close to the target which adjust at only 2.3 percent. This is probably
because off-target leverage firms face higher cost of adjustment. The findings imply
that firm capital structure should be kept close to the target to minimize unnecessary
leverage adjustment costs.

The findings of this study support the relevance of the dynamic trade-off theory.
However, the dynamic adjustment in certain conditions, specifically for near-target and
underleveraged firms, is weak and can be easily dominated by other considerations.
That means that other theories of capital structure might dominate the financing
decisions. If the timing or adverse selection considerations produce benefits that
overcome the costs of deviation, they may dominate the behavior of the firms and the
timing theory or pecking order theory will appear as first order priorities. This is more
likely to occur for firms that are close to the target than firms far from the target and for
firms below the target than firms above the target. These results, though consistent
with the trade-off theory, might be specific to the sample of study and limited by the
controlled variables used to explain leverage. Future studies in other countries are
necessary to validate and generalize the results.

References
Antoniou, A, Guney, Y & Paudyal, K 2008, ‘The determinants of capital structure:
capital market-oriented versus bank-oriented institutions’, Journal of Financial
and Quantitative Analysis, vol. 43, no. 1, pp. 59-92.
Arellano, M & Bond, S 1991, ‘Some tests of specification for panel data: Monte Carlo
evidence and an application to employment equations’, Review of Economic
Studies, vol. 58, pp. 277-297.
Arellano, M & Bover, O 1995, ‘Another look at the instrumental variable estimation of
error-components models’, Journal of Econometrics, vol. 68, no. 1, pp. 29-51.
Baker, M & Wurgler, J 2002, ‘Market timing and capital structure’, The Journal of
Finance, vol. 57, no. 1, pp.1-32.
Baxter, ND 1967, Leverage, risk of ruin and the cost of capital, The Journal of Finance,
vol. 22, no. 3, pp. 395-403.
Booth, L, Aivazian, V, Demirguc-Kunt, A, & Maksimovic, V 2001, ‘Capital structures in
developing countries’, The Journal of Finance, vol. 56, no. 1, pp. 87-130.
Blundell, R & Bond, S 1998, ‘Initial conditions and moment restrictions in dynamic
panel data models’, Journal of Econometrics, vol. 87, no. 1, pp.115-143.
Byoun, S 2008, ‘How and when do firms adjust their capital structures toward
targets?’, The Journal of Finance, vol. 63, no. 6, pp. 3069-3096.
Dang, V A, Kim, M, & Shin, Y 2012, ‘Asymmetric capital structure adjustments: New
evidence from dynamic panel threshold models’, Journal of Empirical Finance,
vol. 19, no. 4, pp. 465-482.
de Miguel, A & Pindado, J 2001, ‘Determinants of capital structure: New evidence from
Spanish panel data’, Journal of Corporate Finance, vol. 7, no. 1, pp.77-99.
Drobetz, W & Wanzenried, G 2006, ‘What determines the speed of adjustment to the
target capital structure?’’ Applied Financial Economics, vol. 16, no. 13, pp. 941-
958.
Elsas, R & Florysiak, D 2011, ‘Heterogeneity in the speed of adjustment toward target
leverage’, International Review of Finance, vol. 11, no. 2, pp. 181-211.
Fama, EF & French, KR 2002, ‘Testing trade-off and Pecking order predictions about
dividends and debt’, The Review of Financial Studies, vol. 15, no. 1, pp. 1-33.

109
 Abdeljawad, Mat-Nor, Ibrahim & Abdul-Rahim
Fischer, EO, Heinkel, R & Zechner, J 1989, ‘Dynamic capital structure choice: theory
and tests’, The Journal of Finance, vol. 44, no. 1, pp.19-40.
Flannery, MJ & Hankins, KW 2007, ‘A theory of capital structure adjustment speed’,
Working Paper, University of Florida, United States.
Flannery, MJ & Rangan, KP 2006, ‘Partial adjustment toward target capital structures’,
Journal of Financial Economics, vol. 79, no. 3, pp. 469-506.
Frank, MZ & Goyal, VK 2007, ‘Trade-off and Pecking order theories of debt’, SSRN
eLibrary, viewed 2 February 2008, <http://ssrn.com/abstract=670543>
Hovakimian, A 2006, ‘Are observed capital structures determined by equity market
timing?’, Journal of Financial and Quantitative analysis, vol. 41, no. 1, pp. 221-
243.
Hovakimian, A & Guangzhong, L 2011, ‘In search of conclusive evidence: how to test
for adjustment to target capital structure’, Journal of Corporate Finance, vol. 17,
no. 1, pp. 33-44.
Huang, R & Ritter, JR 2009, ‘Testing theories of capital structure and estimating the
speed of adjustment’, Journal of Financial and Quantitative Analysis, vol. 44, no.
2, pp.237-271.
Hussain, MI 2005, ‘Determination of capital structure and prediction of corporate
financial distress’, PhD Thesis, University Putra Malaysia, Malaysia.
Iliev, P & Welch, I 2010, ‘Reconciling estimates of the speed of adjustment of leverage
ratios’, SSRN eLibrary, viewed 8 November 2010, <http://ssrn.com/abstract=
1542691>
Jensen, MC 1986, ‘Agency costs of free cash flow, corporate finance, and takeovers’,
The American Economic Review, vol. 76, no. 2, pp.323-329.
Jensen, MC & Meckling, WH 1976, ‘Theory of the firm: managerial behavior, agency
costs and ownership structure’, Journal of Financial Economics, vol. 3, no. 4,
pp.305-360.
Kraus, A & Litzenberger, RH 1973, ‘A state-preference model of optimal financial
leverage’, The Journal of Finance, vol. 28, no. 4, pp. 911-922.
Lemmon, ML, Roberts, MR, & Zender, JF 2008, ‘Back to the beginning: persistence
and the cross section of corporate capital structure’, The Journal of Finance, vol.
63, no. 4, pp.1575-1608.
Myers, SC & Majluf NS 1984, ‘Corporate financing and investment decisions when
firms have information that investors do not have’, Journal of Financial
Economics, vol. 13, pp.187-221.
Myers, SC 2003, ‘Chapter 4 Financing of corporations, in Constantinides, GM, Harris,
M and Stulz, RM (ed.), Handbook of the Economics of Finance, 1st ed, Elsevier,
Philadelphia, vol. 1, no. 1, pp. 215-253,.
Ozkan, A 2001, ‘Determinants of capital structure and adjustment to long run target:
evidence from UK company panel data’, Journal of Business Finance &
Accounting, vol. 28, no. 1, pp. 175-198.
Rajan, RG, & Zingales, L. 1995, ‘What do we know about capital structure? Some
evidence from international data’, The Journal of Finance, vol. 50, no. 5, pp.
1421-1460.
Roodman, D. 2006, ‘How to do xtabond2: an introduction to "difference" and "system"
GMM in Stata’, The Stata Journal, vol. 9, no. 1, pp. 86-136.
Windmeijer, F 2005, ‘A finite sample correction for the variance of linear efficient two-
step GMM estimators’, Journal of Econometrics, vol. 126, no. 1, pp. 25-51.

110