Hypothesis 5: Leverage (Debt/Equity) of REIT and REF may play a key role to determine Return earnings
3. Research Methodology 1 Data and Variables
3.2. Estimation Strategy
3.2.3. Robustness Check strategy
3.2.3.1. Dropped of stock exchange listing variable and excluding data from listed REIT and REF companies
It is stated that our data is comprised with non-listed REIT and REF as well as listed REIT and REF (data based on whole population) where number of stock exchange listed REIT and REF are very few (5 only) but they have 79 observations out of 795 observations consecutively. It is documented that the mean of income before tax-Assets ratio (ROA) for stock exchange listed REIT and REF is 0.057691 and the mean is 0.0208173 for non-listed REIT and REF in same variable (ROA). Some may be worried on selection bias between listed and non-listed firms and
come out as biaseddue to selection or unobservable variable bias. It is also mentionable that in the correlation test, stock exchange listing variable is not found highly correlated with other explanatory variables. To partially check this selection bias, the dummy for being stock exchange listed variable will be used as the dependent variable and all other important characters like REIT (dummy variable), Lagged value of ROA, size of the company, leverage, capital structure, CR-REIT (dummy variable) will be used as explanatory variables. If insignificant coefficients would be got then less concern on selection bias based on the observables. Again, if significant coefficients would be got for some variables then stock exchange listing variable would be dropped in the model and stock exchange listed REIT and REF data have been excluded and run the regression to check the previous results authenticity.
3.2.3.2. Robustness check through excluding high Diversity in data set
Liability_ assets ratio, (the capital structure of a company) is an important variable in my researh. It is observed that the average value of Liability_ assets ratio among the companies is 0.3243997 where mean of Liability_ assets ratio in the case of REIT is higher than REF companies (0.5974769 Vs 0.1624141). Again, it is also observed that there are two REIT companies whose observational data on Liability_ assets ratio in year 2009 and 2012 are above 6 (10.803 and 7.144 consequitively). It is mentionable that according to basic accounting principle, liability and equity is equal to assets and assets is higher than liability. But when a company does not perform well (occuring huge losses) for several years and then liability may go above assets. This situation is referrd as insolvency situation of the company when the company become failed to repay all the liabilities and its own capital (shareholder equity) shows negative sign. As in this research observational data set are based on total population of externally audited REIT and REF companies, some companies are found in the abovementioned condition specially some of the REIT companies data have fallen in this category. It is also mentionable that after excluding the two observations which posses high Liability_ assets ratio, the average value of Liability_ assets ratio has been reduced at 0.3025863. Considering the control of high diversity in Liability_ assets ratio, I exclude the abovementioned two
3.2.3.3. Robustness Check strategy- Global Financial Crisis (GFC)
It is also mentionable that over the sampling period (2001-2018), no subjects (company) has been changed as its operational status at all in this study as for example no REIT has been transformed (no change) to REF or no REF has been transformed (no change) to REIT, no Non- listed REIT has been transformed to listed REIT or no listed REIT has been transformed to non listed REIT, across observational time period. Therefore, a fixed effects model may not work at all.
Again, Difference in Difference is not possible due to REIT has been established in 2001 and before of 2001 there was no REIT company. Parallel trend assumption was not validated in this case. It is also mentionable that considering the observational data size, Propensity Score
Matching (PSM) technique is not possible to apply in this research.
It is documented that Global Financial Crisis (GFC) had been started from real estate sector (financial turmoil began with the bankruptcy of Lehman Brothers on 15 September 2008) and it were spread to US as well as global economy the global capitalist system was threatened including Korean financial market (Korea Economic Institute and Korea Institute for International Economic Policy, 2009). Given the region’s large trade volume and its financial integration with the rest of the world, Korea has been one of the Asian nations most severely hit by the global financial crisis (GFC) in 2008. Investors’ views on the Korean economy deteriorated as global deleveraging intensified that affected the foreign exchange markets as foreigners began to repatriate their funds out of Korean financial markets. As of the end of November 2008, the Korean won had depreciated by over 25.4 percent in dollar terms, the largest fall among major Asian countries excluding Turkey. The stock price collapsed by 27.2 percent during the same period (Kim, K. S., & Chey, H. K. (2010) and equity markets have followed a similar pattern, with the benchmark KOSPI 200 index falling to a low of 948 points in November 2008 (from 2,054 points a year before) before regaining to current levels at 1,684 points (Lim, Jamus 2010).
Some studies documented that after the crisis, the effect of GFC has been continued for several years and real estate financial sector had not been performed as pre GFC period (Senturk, Fatih 2016). To consider this issue, I want to check any shock present or not after the GFC in the financial market thereof. I create After Shocks of GFC variable, a dichotomous/binary variable which represents the unobservable shocks responsible for continual effect after GFC (Year 2010-2018) equal to 1 and Zero otherwise (Year 2001-2009).