“Nexus between financial innovations, remittances and credit performance: Evidence from augmented ARDL and nonlinear ARDL”

The motivation for this study is to assess the impact of financial innovation and re- mittances on bank-based financial institutions’ credit performance in Bangladesh for the period 1981–2019. The study applies augmented ARDL (AARDL) and nonlinear ARDL (NARDL) to identify both long-run and short-run effects and directional causality by performing non-granger casualty tests. AARDL confirms the presence of a long-run association between financial innovation, remittance, trade openness, FDI, and credit performance, which is measured by non-performing loans. In the long run, financial innovation and FDI volatility expose a positive link with NPLs, but remittance inflows and trade openness establish a negative association. Asymmetry shocks in financial innovation reveal a positive relationship with credit performance. In contrast, the asymmetric shock of remittance and trade openness unveil a negative tie to credit performance, especially in the long run. Furthermore, directional causality provides evidence to support a feedback hypothesis explaining causality between financial innovation and credit performance, as well as remittance inflows and credit performance. These findings suggest that credit performance is guided by future de- velopment in remittances and financial innovation; thus, closer attention from policymakers and financial experts is persistent to capitalize or mitigate the impact of the financial system.


INTRODUCTION
The effect of non-performing loans on financial systems is widely used at an international forum. Financial institutions' performance and financial sustainability are adversely affected by the increase of NPLs in the financial system. Therefore, financial institutions repetitively tried to minimize the conversation rate of NPLs arising from the creation of their assets. Furthermore, over the past decades, several reforms and projects have been carried out aimed at reforming banking and financial structures in numerous countries. The enhanced quality of banks' reserves is due to faster growth in nominal income and credit, expanded financial inclusion, as well as the attempts of supervisory authorities to enhance the control of banks' credit risk and underwriting activities. In recent years, the non-performing loan -fundamental macro nexus have attracted immense interest among scholars (Dimitrios et  Credit mobilization is one way to create assets for financial institutions in the financial system, especially bank-based institutions.

LITERATURE REVIEW AND HYPOTHESES DEVELOPMENT
For the last decades, NPLs have become troublesome and one of the most critical issues globally because of banks' profitability and liquidity hamper. Overall economic growth is negatively affected by the increased amount of non-performing loans (Klein, 2013). NPLs respond to macroeconomic, bank-specific, and other global factors. Moreover, wrong customer selection, intense competition, poor management, lack of supporting facilities, poor cash flow, delayed disbursement of funds, and absence of proper monitoring are the gener-al reasons behind the NPLs. Besides, fundamental macro factors such as lower GDP, unemployment rate (Jiménez & Saurina, 2005), inflation rate, interest rate, and the exchange rate (D. Louzis et al., 2010) can be considered as the root causes of NPLs. Furthermore, bank-specific factors (credit growth, bank size, risk-taking, return on asset, return on equity, solvency, inefficiency, liquidity) and global factors, NPLs arise (Boudriga et al., 2009). In addition, macroeconomic and bank-specific variables affect non-performing loans ( In European countries, NPLs are found to increase when unemployment rises, the exchange rate depreciates, and inflation is high (Klein, 2013; D. P. Louzis et al., 2012;Škarica, 2014;Turan & Koskija, 2014). However, opposite findings are found in Shingjergji (2013). An insignificant relationship has been found between GDP, interest rate, inflation, exchange rate, and NPLs (Haniifah, 2015). In a general sense, GDP and NPLs should be strongly related because higher GDP mostly expresses a country's robust economic condition. Non-performing loans cannot exist too much in countries with healthy economic conditions (Ghosh, 2015;Rajha, 2016 Figure 1 represents the possible causality between financial inclusion, stock market development, and foreign capital from a theoretical perspective.

METHODS
The study utilizes annualized time series data for the period 1981-2019. As a dependent variable, non-performing loans are measured by the ratio of NPLs to gross domestic product (Chen et al., 2018). Because of broad regional economic growth disparities, non-performing loans of commercial banks can be affected by regional economic efficiency. Therefore, the non-performing credit ratio is preferred to calculate non-performing commercial bank loans to avoid economic growth. Remittances are measured by the per capita remittance received: Personal remittances comprise personal transfers and compensation of employees. Personal transfers consist of all current transfers in cash or in-kind made or received by residents. It is expected that continuous inflows of remittances in the economy can reduce the trend of increasing non-performing loans in the economy. The expected sign of remittance (R) is negative.
Now the asymmetric shocks derived from equations 4-7 are incorporated in equation (3) and reproduced in a nonlinear framework, which is as follows:

RESULTS
According to the existing literature, empirical estimation using time series data has helped to determine the order of variable integration. In line with literature, the study performs a stationary test by applying the ADF test proposed by Dickey and Fuller (1979), P-P test proposed by Phillips and Perron (1988), KPSS test proposed by Kwiatkowski et al. (1992). The results of unit root tests are displayed in Table 3. The study findings establish a mixed order of integration, indicating that variables are either stationary at a level I(0) or after the first difference I(1); no variables expose stationarity after the second difference I(2).  Table 4.
The results of the long-run cointegration test are reported in Table 5. The study performs nine cointegration equations to ascertain possible long-run association based on NPL measures in the equation and proxy of financial innovation. Because the test statistics of F overall , t DV , and F IDV are statistically significant at a 1% level of significance, these findings suggest the presence of long-run cointegration between financial innovation, remittance, trade openness, FDI volatility, and non-performing loans.
Since the cointegration test of AARDL confirms long-run cointegration in the empiri-   Table 6.
The following section investigates asymmetric effects of financial innovation, remittance, trade openness, and FDI volatility on NPLs by performing a nonlinear ARDL equation. Table 7 exhibits the results of nonlinear models consisting of a six-panel representation.    Note: ***, **, and * indicate significance levels at 1%, 5%, and 10%, respectively.   Note: The subscripts of a/b/c specify the level of significance at 1%, 5%, and 10%, respectively. The next section deals with gauging the directional causality by performing equation (8), and the results are displayed in Table 8. Causality results for NPL for private commercial banks are reported in a panel A; results for state-owned bank are displayed in panel B, and for foreign commercial banks -in panel C.

DISCUSSION
The results of the impact of financial innovation on NPLs establish a positive link between them.
These findings indicate that the excessive financial innovation adaptation and diffusion adversely cause detreating the financial system's loan performance. Referring to the magnitudes of financial innovation, it is apparent that state-owned banks' loan performance in Bangladesh is more intense compared to other segments of bank operation. This finding advocates that bank-based financial institutions must behave rationally by understanding the significance of loan performance and possible consequences of excessive innovativeness in the credit portfolio.
FDI volatility and NPLs nexus reveal a positive association, indicating that the performance of FDI inflow variation in the economy determines the credit performance in the financial system. The evidence is that accumulating long-term capital for investment FDI is positioned on the top because FDI inflows accelerate domestic production activities and promote domestic enterprise expansion. Thus, any innovation in FDI, especially adverse shocks, directly injects a negative wave to-wards domestic firms, and their performance hinders FDI fluctuation. Eventually, the firm's loan performance to financial institutions hampered at large.
The role of remittances in Bangladesh's economy is well appreciated. In the case of overall loan performance s, remittance inflows enhance credit performance by decreasing the propensity to become a credit defaulter. The impact of remittance inflows on credit performance is positively associated, implying that remittances negatively cause NPLs in all segments of bank-based financial institutions. Moreover, in terms of remittance effects, private commercial banks are getting more assistance for decreasing the number of credit defaulters compared to state-owned and foreign commercial banks' operations in Bangladesh.
Trade openness augments aggregate output by expanding business activities by allowing them to participate in the international arena. Trade openness brings a positive wave into the economy, and financial institutions are not isolated from that wave. Referring to model findings, it is noticeable that trade openness helps financial institutions recover their loan performance, which ingresses that negative relationship between trade openness and NPLs.
The error correction term (ECT) represents the speed of adjustment towards long-run equilibrium, which is negative in sign and statistically significant. Table 6 shows that the error correction term (ECT-1) of each specified model is negative and statistically significant at a 1% significance  level. The impact of financial innovation on NPLs is positive in most cases; however, only a few coefficients are statistically significant. Empirical models pass with several diagnostic tests, and their results are reported in Panel C of Table 6. Study findings reveal that all the empirically tested models are free from serial correlation, residuals of error are normally distributed, the issues of heteroskadacity are missing, and internal consistency and stability test confirms models capacity to produce unbiased estimation.
Asymmetric cointegration is evaluated by performing three tests: Fpss, Wpss, and tBDM, and their test statistics are displayed in Panel A (Table  7). All the test statistics are statistically significant at a 1% level, suggesting the rejection of the null hypothesis: "No asymmetric cointegration". This finding suggests that asymmetric cointegration prevails between financial innovation, remittance, trade openness, FDI volatility and NPLs.
In the long run, the asymmetric shocks that are positive and negative in financial innovation expose a positive linkage with NPLs. All the coefficients are statistically significant at a 1% or 5% significance levels. These findings suggest that the role of financial innovation is critical for the banks' credit performance. According to estimations, positive growth can intensify the current state of credit performance degradation and control in financial innovation to manage the present state of credit performance. Moreover, the magnitudes of positive innovation produce more intense than negative shocks in the financial system. Thus, the strategy formulation for credit expansion seeks special attention to gauge financial innovation's role, since it plays a meaningful role in financial development. The asymmetric shocks of FDI volatility establish a positive association with NPLs in the long run, and all the coefficients appear statistically significant at a 1% or 5% levels. A positive association suggests that stability in receiving FDI helps capital accumulation and aggregate productivity expansion, allowing growth in domestic firms. According to the coefficients of asymmetric shocks, i.e., positive and negative innovation, it is palpable that reducing the present state of credit situation can be improved by reducing FDI volatility, implying that stability in FDI inflows lessens the effects of volatility, which will bring positive wave in credit performance in the end. Positive shocks in remittance inflows establish negative associations with NPLs, and negative shocks in remittance inflows expose a positive association with NPLs. These findings suggest that continual inflows of remittances in the economy emerged as an alternative channel for money supply and support for capital formation.
Finally, the asymmetric effects of trade openness on NPLs reveal asymmetric shocks in trade openness and establish negative links with NPLs. Study findings suggest that domestic trade expansion allows firms' earning capacity and increases their credit payment capacity for credit approval from financial institutions. A negative tie between negative shocks and the state of NPLs reveals that decreasing the possibility of domestic trade expansion intensifies the deterioration of credit performance, indicating firms' incapacity for paying credit payment and becoming defaulter eventually.
The results of the short-run model estimation are displayed in Panel C of Table 7. The coefficient of the error term reveals a negative sign and is statistically significant at a 1% level. These findings suggest the speed of long-run convergence due to shocks in the short run, which refers to the shortrun asymmetric impact of financial innovation, remittances, trade openness, and FDI volatility on NPLs. Study findings establish a negative linkage between proxies of financial innovation and NPLs, but in terms of magnitudes, financial innovation is not as critical as appeared in the long run; however, all the coefficients are statistically significant. Regarding the remittance asymmetric shock, only positive variations expose a negative tie with NPLs, and all the coefficients are statistically significant at 1%. Furthermore, the asymmetric impact of trade openness and FDI volatility establish statistically insignificant results in most cases.
Symmetry results for the long term are reported in panel D of  (Table 8). On the other hand, a bidirectional causality between FDI volatility and NPLs is established in Panels B and C. The findings suggest that domestic firm credit performance are largely dependent on macrofundamentals; hence, financial policymakers should emphasize the impact of macroeconomic movements on the implementation of financial policies.

CONCLUSION
The aim of this study is to gauge the effects of financial innovation, remittances, trade openness, and FDI volatility on credit performance in Bangladesh's financial system for the period 1981-2019. An empirical model estimation with AARDL reveals a positive association between financial innovation and NPLs in all banks in Bangladesh. Considering financial innovation elasticities on NPLs, it is evident that state-owned banks show more vulnerability in the process of detreating credit performance, compared to private and foreign banks in the financial system. This finding suggests that state-owned banks must make considerable efforts to understand the ultimate impact of financial innovation while deciding to adopt and disseminate financial innovation in their business models. Furthermore, a positive link reveals FDI volatility to NPLs, indicating that scarcity in capital adversely affects business operations, thus reducing the capacity to perform credit obligations to the financial institutions. This is because FDI inflows are critically significant for supplying capital and smooth business operations.
The results of non-linear ARDL test ascertain the presence of the asymmetric cointegration between financial innovation, remittances, trade openness, FDI volatility and NPLs. Moreover, financial innovation's asymmetric shocks expose a negative linkage with credit performance in all three model estimations. However, in terms of elasticities of asymmetric shocks, the study found that positive shocks in financial innovation are more prominent than negative variations. This fining explains that lessening the present state of credit performance, that is the reduction of credit defaults in the financial system, it is immensely significant to establish control over the speed of financial innovation diffusions in the system. The asymmetric effects of remittance and trade openness reveal a positive link with credit performance in the financial system. The role of remittances and trade openness is critically significant because of money supply management, financial efficiency, and capital accumulation from the expansion of domestic business. Finally, continual inflows of FDI are critical for the financial system, since longterm investment and capital asset mobilization increase aggregate productivity and create domestic enterprises' scope for glut earnings.
In summary, the findings reveal that excessive financial innovations play a detrimental role in bankbased financial institutions' credit performance, both in the long and short run. Thus, financial institutions have to be very cautious when offering innovative credit products, though financial innovation helps enhance financial performance; however, adverse effects can cause credit performance with a negative note. The expansion of domestic output accelerates the credit scope for the financial institutions and, at the same time, boosts credit performance. Therefore, contraction economic policies can adversely cause trade liberalization and, ultimately, contribute to credit de-regularization in the financial system.