Sources of Fluctuations in Islamic, U.S., EU, and Asia Equity Markets: The Roles of Economic Uncertainty, Interest Rates, and Stock Indexes

ABSTRACT

This article analyzes the economic and financial sources of fluctuations among the U.S. federal funds rates, the U.S. economic policy uncertainty, and the indices of the U.S., European, Asian, and Islamic stock markets. The impulse response analysis shows that the U.S. economic policy uncertainty shocks have significant and negative effects unanimously on the U.S., European, Asian, and Islamic stock markets. A contractionary monetary policy shock, in terms of a higher federal funds rate, has also a statistically significant and negative effect on all of the stock markets. The variance decomposition results indicate that the Islamic stock index is mainly affected by the U.S. stock index shock, thus negating its dichotomy hypothesis. The U.S. economic uncertainty shock explains an important portion of fluctuations for all four stock indices. The degree of synchronization between the EU stock market and other markets has weakened after the U.S. financial crisis.

KEY WORDS:

• global Islamic stock markets
• SVEC model
• synchronization
• economic uncertainty

Notes

1. The reader is referred to Baker et al. (2013) and www.policyuncertainty.com for further details regarding the construction of the economic policy uncertainty indexes for the United States and other countries.

2. There is also a growing literature on Islamic banks (see for example,Hesse et al. 2008; Abd Rahman 2010; Cihak and Hesse 2010; and Khan 2010). Sole (2007) also presents a “good” review of how Islamic banks have become increasingly more integrated in the conventional banking system.

3. The unit-root test and the cointegration test results are available upon request.

4. For example, Ayalew et al. (2012) showed that the Hannan-Quinn Criterion (HQC) is the best with the large sample data. In our case, the sample size is 3,601 so we use HQC in choosing optimal lags for VECM.

5. With these different time zones, the short-run ordering should have been ASIA→EU→US→ISLAM. However, with this ordering, the U.S. effects on the other markets become very small. Thus, we use the data with time “t” for the U.S. and Islamic variables, while “t+1” for the Asian and EU variables.

6. Even though our identifying restrictions are based on the informational lags for the short-run impact matrix and economic theories for the long-run impact matrix, we also perform two different types of causality tests: the Granger causality test and the instantaneous causality test. The test results are generally consistent with our identifying restrictions in the following ways. First, we imposed long-run restrictions such that EPU does not affect all other variables in the long run (or EPU is determined by all other variables in the long run, the relationship is given as cointegrating relationship). If Granger-causality is meant to be used for the long-run identifying restrictions, our long-run restriction on EPU is consistent with Granger-causality result. Second, our short-run identifying restrictions between the different stock indexes are U.S→ISLAM→ASIA→EU and the instantaneous causality test reveals that there is no instantaneous causality between 1) Asia, 2) EU and the other variables. Thus, there seems to be no conflict between our identifying restriction and the results from causality tests.

7. The detailed impulse response results for pre- and post-financial crisis periods are available upon request.