Disentangling the relationship between remittances and financial development: evidence from Jamaica

ABSTRACT

This article examines the relationship between remittances and financial development in Jamaica using annual data from 1976 to 2016. We apply Principal Component Analysis to construct an index which captures the dimensions of different indicators of financial development. Using an ARDL approach to construct an error-correction model and the Toda-Yamamoto test to examine causality, we distinguish between the long-run and short-run dynamic linkages between remittances and financial development. Utilizing several financial development models to account for the various empirical relationships in the literature, we find that remittances promote financial development in the long run, while substituting for financial development in the short run. Additionally, the long-run effect is larger, and there is a lag before the short-run effects are realized. These findings suggest that remittances may have different roles in the process of economic development. Thus, differentiating long-run and short-run policies are likely to be important to harness the developmental impact of remittances.

KEYWORDS:

• Remittances
• financial development
• ARDL
• principal component analysis
• Jamaica

Disclosure statement

No potential conflict of interest was reported by the authors.

Notes

1. Remittances were found to be more stable than foreign direct investment (FDI) flows and other capital flows during financial crises (World Bank 2015b). Remittances to developing countries fell by 3.7% following the financial crisis in 2008, while FDI inflows fell by 31.4% (World Bank 2019).

2. A number of studies examine single countries based on micro-data. However, these investigate short-run impacts and therefore cannot address the long-run effects.

3. Value added by the financial sector has increased, in comparison to generally declining trends in other sectors (Statistical Institute of Jamaica 2018).

4. Market power may distort financial intermediation (Moore and Craigwell 2002).

5. Issahaku, Abor, and Harvey (2017) construct a measure for banking sector development in their panel data study on remittances.

6. The theoretical model of Brown and Carmignani (2015) incorporates financial market imperfections, but they relate this to a non-linear effect of remittances.

7. Remittances to Jamaica were also less volatile over the past decade, as measured by the 3-year rolling window standard deviation.

8. These weights are standardized so that they sum to 1. For further discussion, see Ang and McKibbin (2007).

9. Given our limited dataset, asymptotic critical values may be inappropriate for the analysis. Narayan (2005) provides small sample critical values, which we use for the test.

10. We also find that the control variables are not integrated of an order greater than one. These results are available upon request.

11. The columns in Table 6 correspond to those presented in Table 5.

12. The regression specification in Column (5) of Table 5 does not pass the Ramsey test for misspecification. This may be due to the presence of nonlinearities. As this is not the main focus of the study, we do not explore nonlinear regressors. However, all other diagnostic tests are passed.

13. In a study on remittances and growth for Caribbean countries, Deonanan and Ramkissoon (2018) find remittances to be unaffected by financial development in Jamaica.

14. More specifically, in cointegrated systems, causality at frequency zero is considered to be ‘long-run causality’ (Breitung and Candelon 2006).

15. We can reject the null for frequencies less than 1.33, which corresponds to cycles greater than 4.7 years.