The Possibility of Manipulated IPO Underpricing and Post-IPO Stock Return : Empirical Study on Jakarta Stock Exchange 2009-2012

This research pursue the uniqueness of Indonesia Initial Public Offering (IPO) that can not be accurately explained by previous studies conducted in developed countries. Theories and empirical findings on IPO cover many issues such as asymmetric information (Baron, 1982), shareholders wealth maximization (Aggarwal et al, 2002), asymmetric information and signaling by high quality firms (Ritter and Welch, 2002), decision of percentage of shares to be offered (Zheng et al, 2005), high initial return but low long-run return (Georgen, 2007), and also many other empirical reports. However, this research offers modified approaches to adopt specific Indonesia context. The discussion of Indonesia IPO is started with how the allotment rules for IPO shares are implemented.1 The underwriters were allowed to separate issued shares into two kinds of allotments, namely fixed allotment which is a much larger proportion of around 98% of the total new shares, and pooling allotment which is only 2% of the total shares. The fixed allotment shares are usually offered to institutional investors during the book building before the offering period, and the shares of pooling allotment The Possibility of Manipulated IPO Underpricing and Post-IPO Stock Return: Empirical Study on Jakarta Stock Exchange 2009-2012


Introduction
This research pursue the uniqueness of Indonesia Initial Public Offering (IPO) that can not be accurately explained by previous studies conducted in developed countries.Theories and empirical findings on IPO cover many issues such as asymmetric information (Baron, 1982), shareholders wealth maximization (Aggarwal et al, 2002), asymmetric information and signaling by high quality firms (Ritter and Welch, 2002), decision of percentage of shares to be offered (Zheng et al, 2005), high initial return but low long-run return (Georgen, 2007), and also many other empirical reports.However, this research offers modified approaches to adopt specific Indonesia context.
The discussion of Indonesia IPO is started with how the allotment rules for IPO shares are implemented. 1The underwriters were allowed to separate issued shares into two kinds of allotments, namely fixed allotment which is a much larger proportion of around 98% of the total new shares, and pooling allotment which is only 2% of the total shares.The fixed allotment shares are usually offered to institutional investors during the book building before the offering period, and the shares of pooling allotment Ritter and Welch (2002) explain there are two types of IPO firms, namely higher quality firms and lower quality firms.This research propose the third type, namely bad IPO firms which manipulate and force IPO underpricing.Bad IPO firms are subset of lower quality IPO firms that force false signal as higher quality firms.The false signal was hidden by managing post-IPO trading.Trading management are indirectly funded by using balance sheet cash.Hypothesis testing with the empirical model 1 was to confirm the role of CashRatio as the moderating variable that interact DER to affect IPO underpricing which originally was not.The findings support the predictions that interactive variable DER*CashRatio affect IPO underpricing.A managed trading had a non negative profits constraint so that selective post-IPO trading was conducted to cause trading imbalance observable as skewed trading volume (Skewness).Subsequent tests with the empirical model 2 was to confirm the role of Skewness as the moderating variable that interact VolRatio to affect post-IPO stock return (RGM) which originally was not.The findings support the predictions that interactive variable LnVolRatio*Skew affect RGM.Both findings confirm this research predictions on the possibility of manipulated IPO trading in Indonesia IPO 2009-2012.would be allocated to retail investors.Fixed allotment would be a guaranteed purchase orders through the underwriters; these are bidding commitments so that the demand for the fixed allotment will be known before the second and final filling to the regulator.
On the contrary, demand by retail investors of pooling allotment shares would only be known after the offering period; it could not be known in advance.
After the offering period, it would also be known whether the number of shares demanded exceed the number of shares allocated on the pooling allotment.At this stage, news of either over-subscribe or under-subscribe will be known by public; which was actually refer to the pooling allotment only.Fixed allotment would not be changed, an order cancellation of the fixed allotment was not an option.
On the other side, new demands by institutional investor was also possible.However, the underwriter would only fulfil the new demands by selling reserves from the fixed allotment if the underwriter had planned to do so.Fulfilment of new demands of institutional investors from the pooling allotment is not preferable because of the risks of decreasing number of public investors. 2here are two expected good news of IPOs : (1) Over-subscribe, and (2) IPO underpricing. 2 Logically, an over-subscribe IPO would be followed by an IPO underpricing.But it is not always the case.
It is also important to note that unlike in New York Stock Exchange or other developed stock exchanges, there is no market maker in Indonesia Stock Exchange (IDX).
As explained, pooling account is merely small portion of the IPO.This is the root of opportunity for manipulation why over-subscription along with IPO underpricing can be created as signal of successful IPOs.
If manipulated IPO trading exists, then the funding is needed.There should be role of founder shareholders to force top management of IPO firms to allow indirect utilization of balance sheet cash.Other than this, subsequent manipulation in the form of managed post-IPO trading should be done to hide the false signal of IPO underpricing.But there is a constraint of non-zero profits during the market operations which create certain trading pattern.
With regard to the above preliminary observations, this research name the IPO firms which conducted manipulated trading as bad IPO firms.
The research problem is whether the presence of bad IPO firms can be explained by factors related to balance sheet cash and post-IPO trading volume pattern.
There are two research questions: (1) What factors and whether the ratio of the balance sheet cash to the IPO proceed affect IPO underpricing in Indonesia, and (2) What factors and whether the volume trading skewness affect post-IPO stock return.
The objective of this study is to find the answer of the possible existence of bad IPO firms that manipulated IPO underpricing and managed post-IPO trading.
Basically, the newness and main contributions of this research are : 1).Modification of the IPO firms concept of Ritter and Welch (2002) and raising the concept of so called bad IPO firms, 2).Application of moderating variable related to balance sheet cash that explain IPO underpricing, 3).Application of moderating variable related to post-IPO trading volume that explain post-IPO stock return, 4).A method that bad IPO firms can be empirically observed and explained.

Literature Review
Previous reports of IPO studies consider many important things which are more applicable to the developed capital markets.Some of previous IPO studies are the explanation of percentage of shares to be offered, IPO pricing, and future stock performance.Zheng et al (2005) state that during an IPO, several important factors are considered important which include the percentage of shares offered, IPO pricing, and IPO underpricing.Other than these, long-run trading liquidity is one of the founder shareholders objectives.Georgen (2007) reemphasize the past empirical evidences of two main anomalies on the IPO, notably high initial return but poor long-run return.
Other objective of IPO underpricing is the long-run founder shareholders wealth maximization as explained by Aggarwal et al (2002), and Ang and Brau (2003) as cited by Zheng et al (2005).
Empirical studies also provide evidences of IPO underpricing according with finance literature that provide theoretical explanation including asymmetric information as reported by Baron (1982), Rock (1986), Welch (1989), Grinblatt and Hwang (1989), Faulhaber (1989), andBalvers et al (1993).IPO firms deliberately underprice their shares to create signal of a good quality.However, to create signals the IPO firms should bear the cost of underpricing; they sell the shares lower than fair market price.
Booth and Chua (1996) explain IPO underpricing will induce investors to provide information to the IPO firms and eventually will lead to IPO over-subscribing.Ritter and Welch (2002) explain asymmetric information on IPO that higher quality firms deliberately underprice their shares to deter lower quality firms from imitating their actions.Theoretical framework of this research is built and based on the IPO signaling theory.But modifications are developed to accommodate the Indonesia context.The main difference of arguments of this research compared with Ritter and Welch (2002) is lower quality IPO firms may be able to imitate signal by higher quality IPO firms by creating false signal of manipulated IPO underpricing and managed post-IPO trading.The false signal might not be known by investors during the period of price management and operations.Aggarwal and Wu (2006) explain variety ways of stock price manipulation that can also be taken by corporate insiders.It is explained how insiders may manage stock price and trading volume with the objective of creating trading profits.
Chen et al ( 2001) report the relation of stock market return, trading volume, and volatility in nine major stock markets and found that trading volume provides information to the stock index.
The arguments of Ritter and Welch (2002) on IPO underpricing as signaling imply: (1) IPO firms hold superior information relative to the investors regarding the future prospects of the firms, (2) There exist lower quality of IPO firms try mimicking the signal of higher quality IPO firms by IPO underpricing but could not to do so because IPO underpricing is costly.
This research build arguments for signaling concept of IPO underpricing as follow: 1.Higher quality IPO firms send true signal and the investors believe it.This is a real IPO underpricing.2. Higher quality IPO firms send true signal.
But investors do not believe it.IPO underpricing may not exist.3. Lower quality IPO firms which can not send signal be honest, there is no IPO underpricing.4. Some lower quality IPO firms send false signal and force IPO underpricing.These types of firms are named bad IPO firms.Bad IPO firms were able manipulating IPO underpricing by utilizing balance sheet cash indirectly.They manipulate the trading by utilizing founder shareholders shares which is much larger than the pooling allotment shares that only 2% of the total IPO shares.Founder shareholders had to manage and negotiate with investors that hold significant fixed allotment shares not to sell their shares in order to avoid stock price down.In Indonesia case, it is possible if IPO float is not large and the respective investors are within the group of founder shareholders.IPO trading manipulation should be continued until post-IPO to hide the false signal.Funding requirements were facing a non negative profits constraint so that selective post-IPO trading activities had to be employed.This constraint may create trading imbalance observable as skewed post-IPO trading volume.If the investors do not know the signal was actually false, then the trading support by bad IPO firms would be minimal.
The presence of bad IPO firms is confirmed when these two conditions are met, namely manipulated IPO underpricing and managed post-IPO trading.Lower quality IPO firms are defined as firms with high debt-to-equity ratio (DER) and/ or negative operating cash flow.A bad IPO firms are defined as a subset of lower quality firms with these following additional arguments: High DER alone may not cause IPO underpricing.But DER moderated by CashRatio will positively affect IPO underpricing.The operational variable is the interactive variable DER*CashRatio to affect IPO underpricing.It is the necessary condition.
VolRatio which is measured as the ratio between first-day trading volume and the mean of post-IPO trading volume may not affect stock return.But VolRatio moderated by post-IPO volume trading skewness may have significant effect to post-IPO stock return.Interactive variable of Volratio*Skew to negatively affect post-IPO stock return.It is the sufficient condition.

Hypotheses of IPO underpricing
IPO underpricing is measured as Initial Return.The main focus of the model 1 is the variable of CashRatio which moderate DER to affect Initial Return.DER should not affect Initial Return; as well as CashRatio should not.But DER moderated by CashRatio should affect Initial Return in order to confirm a manipulated IPO underpricing to exist.Interaction variable of DER and CashRatio, namely DER*CashRatio is predicted to positively affect Initial Return.Mendez (2011) found the higher the floating the lower the underpricing.The higher the IPO underpricing the higher the price-to-earningratio (PER) of the IPO firms ceteris paribus, so that difference of IPO firms PER and sub-sector PER should positively influence the degree of IPO underpricing.Return on equity (ROE) measures the management performance with respect to the founder shareholders equity; the higher the ROE the better the performance and the higher the firm quality.DER should positively affect initial return.Habib and Ljungqvist (2001)

Hypotheses of post-IPO stock return
Higher quality IPO firms let the liquidity (volume), price, and stock return determined by the market post-IPO trading.On the opposite, bad IPO firms have to continue managing and supporting the post-IPO trading for certain trading period with the objective to continue imitating and confirming signal as higher quality firms.A manipulated but smooth trading may be too costly if not followed by many investors participating in the trading.With non negative profits as the constraint, there will be less liquid post-IPO trading volume relative to the first-day IPO trading.Hence, skewed trading volume may exists.The prediction is that post-IPO stock return is affected by volume ratio which is moderated by volume trading skewness.Ibbotson (1975) reported negative effect of initial return to long-term performance.However, in the presence of bad IPO firms initial return is the basis of the return management.This research predict initial return positively affect post-IPO return.
Volume of float should negatively affect post-IPO return to guarantee non negative profits because the trading would be more manage-able to bad IPO firms if the number of stocks issued were smaller.
Pre-IPO balance sheet cash in the form CashRatio had been predicted to affect initial return in model 1.Post-IPO balance sheet cash will also be tested in model 2 whether it affect post-IPO stock return.Schultz and Zaman (1994) show aftermarket support of an IPO is necessary to avoid stock price decline.Robinson et al (2004)

Empirical models
There are two empirical models adopted.Model 1 will be used to test factors that cause IPO underpricing, and model 2 will be used to test factors that cause post-IPO stock return.

Result and Discussion
Descriptive statistics, estimation output model 1, and estimation output model 2 are presented in Table 3, Table 4, and Table 5  A window of 97-trading day is chosen so that the last trading day was in 14 May 2013 of the last IPO in 19 December 2012 IPO. 4   Diagnostic tests. 5est of multicollinearity for model 1 is conducted by using Variance Inflation Factor defined as VIF = 1/(1-ρ 2 ) with a critical value of 10.There is no multicollinearity in the model 1.
Test of heteroscedasticity for model 1 is conducted by using Glejser heteroscedasticity test with a critical value of │+/-1.96│ at α=5%.There is no heteroscedasticity in the model 1.
There is also no multicollinearity on the model 2. Glejser heteroscedasticity test for model 2 provide the highest t-value of 2.72.However, there is no opportunity to take natural logarithm on it because IR value could be negative.Other than this, standardization by taking the square root of negative numbers is not defined.6

Squared R and F-tests.
On the model 1, The R squared is 32.9% so that the variations of the dependent variable are only 32.9% explained by the variations of the independent variables.However, F-test of 4.49 is larger than the critical value of F-table of ~ 2.10 at α=5% (with degree of freedom of 8 for numerator and 72 for denominator) so that the null hypothesis of altogether β 1 = β 2 = β 3 = β 4 = β 5 = β 6 = β 7 = β 8 =0 should be rejected at α=5%.Model 1 is said to be valid.
On the model 2, the R squared is 22.5% so that the variations of the dependent variable are only 22.5% explained the variations of the independent variables; while the other 77.5% are explained by the other factors do not included in the model.However the F-test with a null hypothesis of altogether δ 1 = δ 2 = δ 3 = δ 4 = δ 5 = 0, with the degree-of-freedom of 5 for numerator and 76 for denominator at α=5% give a critical value ~ 2.37.The calculated F of 4.43 is larger than the critical value so that the null hypothesis should be rejected.The model 2 is also valid.The higher the DER*CashRatio the higher the IR (underpricing) the more favorable the IPO firms to the investors.

Analysis of IPO underpricing
In relation with manipulated IPO underpricing, lower quality IPO firms with high DER without high CashRatio would not be able to force IPO underpricing and could not imitate signal as explained by Ritter and Welch (2002).But bad IPO firms were be able to do so by utilizing the capacity of cash which is sufficiently available in the balance sheet.The higher the DER*CashRatio which is in the same direction with the higher the ratio of balance sheet cash to the IPO proceed, the more power and confidence of bad IPO firms to finance the trading against shares from the pooling allotment. 7(b).Vfloat.The t-test for Vfloat is |-2.81| higher than critical value of |-1.96|; the null-hypothesis is rejected at α=5%.The conclusion is Vfloat negatively affect initial return; the higher Vfloat the lower initial return.Recall that the measurement of Vfloat is the natural logarithm of the ratio of the first-day trading volume to the pecentage of IPO floating.The expectation according to the theory is the higher the percentage of IPO float (denominator of the Vfloat measurement) the lower the IPO underpricing.However, the result is the opposite.The possible explanation is in the case of bad IPO firms an IPO underpricing is a manipulation.The offer price was not underpriced and the first-day closing price was pushed upward; these activities were to create trading profits.In this case the higher the IPO float the higher the IPO underpricing in order to maximize manipulated trading profits. 8 The results also imply the higher the first-day trading volume (numerator of Vfloat measurement) the lower the IPO underpricing.In the case of bad IPO firms, any involvement of fixed allotments investors outside the bad IPO firms group would affect manipulation activities.If the trading participation of these outside in- vestors along with pooling allotment investors were low, then the trading profits would be low because outsider participation volume were low.However, it was easier to push the price up to maximize trading profits.If the participation were high which is observable as high firstday trading volume, then the IPO underpricing would be low because it was more difficult to push the price up.However, trading profits of bad IPO firms group might be high because outsider participation volume were high.These are the explanations why Vfloat negatively affect initial return.
(c).DPER.Initial Return is positively affected by DPER at α=5%.Stock Price is determined by the earnings multiple PER and IPO firms earnings.If bad IPO firms exist, then logically they prefer managing a high PER rather than a high earnings.Earnings management is possible.However, it is risky; earnings figure will be recorded as operating performance of bad IPO firms.On the other hand, a high PER is more subjective and determined by the market so that an upward bias is considered as market valuation.
(d).LnSales.The finding shows that LnSales negatively affect initial return.IPO underpricing is a signal (Ritter and Welch, 2002).Arguably, sales is also a signal of quality.The higher the sales the better the firms image.The higher the sales the lesser the degree of IPO underpricing required as signal as good firms; ceteris paribus.The negative causality of LnSales to IPO initital return (underpricing) is as expected.
(e).Dboard.The findings show that DBoard positively affect IPO underpricing.Even if there is no specific differences in treatments of trading in the two listing boards in IDX namely main board and development board, the listing in the main board is still considered more bona fide.Actually, to be listed in the main board in IDX is not a challenging task.Three main criteria are the year of incorporation, subsequent positive earnings, and the number of public investors more than one thousand.When considering IPO underpricing as a signal of a good firm quality, a listing in the main board should be preferable.Bad IPO firms prefer pursuing a main board listing to strengthen their false signal.Hence, the positive effect of Dboard to Initial Return is as expected.

Analysis of post-IPO stock return
Hypothesis testing is conducted with significance level α=5% so that the critical value is +/-1.96 to decide whether each independent variable to affect post-IPO stock return.The estimated t-statistic can be seen in

Combined results of both models
The hypotheses testing of model 1 confirm IPO underpricing is positively affected by DER*CashRatio with CashRatio as the moder-ating variable.
The hypotheses testing of model 2 confirm post-IPO stock return is negatively affected by LnVolRatio*Skew, with Skewness as the moderating variable.
Explanation of DER*CashRatio to initial return is the first or necessary condition to the presence of bad IPO firms which employing the power of CashRatio to manipulate IPO underpricing.Further, the explanation of LnVolRatio*Skew to post-IPO stock return is the second or sufficient condition for the presence of bad IPO firms with skewed post-IPO trading volume to affect stock return.

Conclusion
The position of this research is to propose the concept of the presence of bad IPO firms.The first main finding is bad IPO firms could force IPO underpricing by utilizing the ratio of balance sheet cash to the IPO proceed.The second main finding is non negative managed trading constraint would create skewed post-IPO trading volume which affect post-IPO stock return.The first research questions can be answered that five factors, namely DER*CashRatio, Vfloat, DPER, LnSales, and Dboard affect IPO underpricing.Particularly, CashRatio has significant contribution that interact to debt-to-equity ratio to affect IPO underpricing.The second research question is also can be answered.Two factors, namely LnVolRatio*Skew, and IR affect medium-term IPO stock return.Particularly, Skewness has significant contribution that interact the ratio of first-day IPO trading volume to the mean of post-IPO trading volume to affect post-IPO stock return.With these two main findings, it can be confirmed the possible existence of bad IPO firms that manipulated IPO underpricing and managed post-IPO trading.
Estimation of model 1 provide evidences IPO underpricing are explained by five explanatory variables.But the other three variables do 118 INDONESIAN CAPITAL MARKET REVIEW • VOL.VI • NO.2

The Possibility of Manipulated IPO Underpricing and Post-IPO Stock Return: Empirical Study on Jakarta Stock Exchange 2009-2012
: Sales affect Initial Return.H 1.6.: Age of IPO firm affect Initial Return.H 1.7.: Listing board affect Initial Return.H 1.8.: Five-day market return affect Initial Return.
explain the role of underwriters that the greater the stock retention the lower the probability of required after market price support.

Table 1 .
Variable definition of model 1

Table 2 .
Variable definition of model 2

Table 3 .
Descriptive statistics of raw data With regard to model 1, this research focus on the possible presence of bad IPO firms by investigating the role of CashRatio as the moderating variable.The estimated coefficient for DER*CashRatio is significantly different from zero at α=5%.The conclusion is DER*CashRatio positively explain Initial Return as expected.Other regression had been conducted that DER and CashRatio as two separate independent variables each does not explain Initial Return IR.It is concluded that CashRatio has significant contribution to the causality of DER to Initial Return.DER*CashRatio is the interaction variable that affect IR.Test results of model 1 confirm CashRatio as the moderating variable. (a).DER*CashRatio.

Table 4 .
Model 1 estimation output: Test for IPO underpricing

Table 5
LnCashPostIPO, and RMGM.Two separate independent variables VolRatio and skewness each does not affect stock return.The causal-

INDONESIAN CAPITAL MARKET REVIEW • VOL.VI • NO.2 120Table 5 .
Model 2 estimation output: Test of post-IPO stock return IR.The results show a t-value of 2.48 so that the null hypothesis of δ 2 = 0 is rejected at at α=5%.Initial Return IR positively affect post-IPO stock return; IPO firms which exhibits positive IR will continue delivering positive post-IPO return on the medium-term.In the case of bad IPO firms, the trading management will pursue positive post-IPO return to strengthen the false signal of IPO underpricing as long as non negative constraint is met.(c).Vfloat.The result shows Vfloat do not affect post-IPO stock return.The indirect effect of Vfloat to medium-term stock return, if any, has already been captured by Initial Return IR on the test of model 1.(d).LnCashPostIPO and RMGM.The result shows that LnCashPostIPO does not affect post-IPO stock return as expected.Stock valuation already consider post-IPO cash, there should be no surprise on this variable.The test on RMGM also shows no causality effect of RMGM as the control variable to post-IPO stock return.