Money supply behaviour in emerging economies: a comparative analysis

Abstract

This paper reports new evidence consistent with the post-Keynesian hypothesis of money endogeneity for hitherto unexplored 10 emerging economies. These results were obtained using a vector error correction model to test for long-run and short-run causalities with data from 1996 to 2007. The evidence suggests that money supply is endogenous in five countries, namely China, the Czech Republic, India, Malaysia and Turkey; it is exogenous in Mexico, while there was no causality found in Indonesia, Russia and Taiwan. Thailand showed endogeneity in the long-run causality. Some suggestions are made to explain the mixed results, and we also discuss the limitations arising from our narrow specifications of the money supply and the models.

Keywords:

• post-Keynesian
• endogenous money supply
• vector error correction model

JEL classifications:

• E12
• E51

Notes

∗∗∗Significance at 1% level; critical values are MacKinnon one-sided values; C and T indicate that the constant and time trend, respectively, included in the Phillips–Perron test are significant.

∗Significance at 10% level

∗∗∗significance at 1% level; critical values are MacKinnon one-sided values; C and T indicate that the constant and time trend, respectively, included in the Phillips–Perron test are significant.

∗Significance at the 1% level.

∗∗significance at 5% level

∗∗∗significance at 1% level

∗∗significance at 5% level

∗significance at 10% level

1. In some literature, this is known as credit money, as money is derived from credit supply.

2. Six channels are suggested. These include the narrow credit channel (or bank-lending channel), the broad credit channel (also known as balance sheet channel), interest rate channel, monetarist channel, the wealth effect channel and the exchange rate channel (see Arestis and Sawyer 2006 for details).

3. For a detailed discussion on reforms in Asia, see Ariff and Khalid (2005).

4. The details provided in Table 1 and the conjectures made in this section are based on discussions in a detailed report by the Bank of International Settlement (BIS 2008).

5. As the unit root test used is the Phillips–Perron test which uses the Newey West technique to determine optimal lag length, we used a different optimal lag length criterion, i.e. the SBC for the VAR and VECM.

6. See Table 1, page 151, in Khalid (2006).