Abstract
Cointegration analysis suggests that the buffer stock precautionary model accounts for the optimal reserve demand in nine developing countries located in Asia and Latin America. The corresponding VECMs are further interpolated, using the permanent and transitory innovation decomposition procedure, in order to assess the relative impact of the time series on the convergence to equilibrium after a shock. Finally the (asymmetric) effect on the speed of convergence of positive/negative changes in signal variables—such as the excess reserves of the previous period, relative competitiveness and US monetary stance—is found to be significant, in line with mercantilistic and fear of floating motives for hoarding reserves.
This is a preview of subscription content, log in via an institution to check access.
Similar content being viewed by others
Notes
Their model can be summarized as follows. Reserves follow in their model a Wiener process and, immediately after a restocking, the authorities select the initial level that minimizes total expected costs. The latter have two interrelated stochastic components: (1) the opportunity cost of reserve holdings and (2) the adjustment cost of reserve restocking whenever the latter have reached a lower limit, set here to zero. The adjustment cost stems from the output reduction brought about by the need to generate the balance of payments surplus which will in turn generate the reserve build up. A higher variability implies that reserves are likely to reach their lower bound more often and require costly restockings. The authorities are faced with a standard cost-benefit choice: the larger the stock of reserves, the lower is the expected cost of adjustment and the higher the expected value of the opportunity cost (and vice versa).
Reserve holdings are reduced if the opportunity cost rises and increased whenever the volatility index rises. The coefficient of the value of imports is positive. It reflects the requirements of international trade on the banks’ transactions demand for reserves and is positively related with the country’s openness. We are not imposing here the additional restrictions that b 1 = 0.5 and b 3 = -0.25 even if, as we shall see below, the estimated cointegration coefficients take values that are surprisingly close to the theoretical ones.
The value of test statistics significant at the 5 percent level and the corresponding dates are available from the authors upon request.
The presence of ARCH effects is corroborated by the serial correlation of the squared reserve increments, assessed with the help of Ljung Box Q-statistics.
rt is the opportunity cost of holding reserves, which, in the case of emerging markets, are mostly invested in US assets. It should thus be measured as the spread between a domestic interest rate and a US Treasury bond interest rate. Since the latter is small and tends to vary but little with respect to the domestic rate we quantify the opportunity cost as the emerging market rate. Only short rates are available over the 1985–2006 time span.
The statistics take the following values, where \(\mathop {\sup }\limits_{1 \leqslant t \leqslant T} Q_T^{\left( t \right)} \) is the supremum of the Nyblom statistic \(Q_T^{\left( t \right)} \) defined in Hansen and Johansen (1999, p.315, eq. 20). The 5 percent critical values are tabulated in Dennis (2006).
Argentina
Brazil
Chile
Mexico
Venezuela
Indonesia
Korea
Malaysia
Singapore
\(\mathop {\sup }\limits_{1 \leqslant t \leqslant T} Q_T^{\left( t \right)} \)
0.57
1.20
0.99
0.32
1.45
1.34
2.39
2.65
1.11
5 percent critical values
5.36
4.62
4.69
5.36
3.16
4.62
4.69
5.36
4.69
The short run dynamics are concentrated out prior to performing the recursive estimation. The estimates set out in Table 1 are obtained with the DOLS procedure. As pointed out in Stock and Watson (1993) they are asymptotically equivalent to the Johansen ML estimates tested in this table.
It should be noticed that in the case of Chile and Korea the variability coefficient estimate is also close to the theoretical absolute value posited by the Frenkel and Jovanovic model. The same result is obtained for the interest rate coefficient in Indonesia and Venezuela.
The VAR lag order is similar to the order used in the trace cointegration tests. We base its choice, in each country, on two criteria: (1) the absence of serial correlation of the residuals (using a multivariate LM test for residual correlation originally set forth by Johansen 1995, page 22) and (2) a Wald test for the joint significance of all lagged endogenous VAR variables up to the selected lag.
Details on the Gonzalo Ng procedure implemented in the paper can be found in Cifarelli and Paladino (2007).
We could associate, following Lettau and Ludvigson (2004), the temporary shock with interest rate behavior in Chile. The same reasoning links the three permanent shocks with reserves in Singapore. No clearcut association is possible in the remaining countries, where both permanent and temporary shocks account for the variability of all the variables.
These shifts are not followed by a full adjustment of reserves to their trend value thus generating a transitory component and a temporary cointegrating error. This is not the case of Chile and Singapore, where reserve variation is not affected by the temporary shock, and reserves adjust rapidly to permanent innovations in the remaining variables in the cointegration equation. In these two countries the coefficient of the interest rate could be misleading; it indicates the impact of the permanent component only, which accounts for but a tiny fraction of interest rate variation.
The half-life of a deviation from long run equilibrium is, in these countries, somewhat longer and reaches 9.85 months on average.
Both the Fed fund and the national REERt time series are I(1). Their first differences are I(0) and may be interpreted as “news”.
References
Aizenman J, Marion N (2004) International reserve holdings with sovereign risk and costly tax collection. Econ J 114:569–591
Aizenman J, Lee Y, Rhee Y (2004) International reserves management and capital mobility in a volatile world: policy considerations and a case study of Korea. NBER Working Paper Series n. 10534
Aportela F, Gallego F, García P (2005) Reserves over the transitions to floating: lessons from the developed world. Mimeo, Central Bank of Chile
Bahmani-Oskooee M, Brown F (2002) Demand for international reserves: a review article. Appl Econ 34:1209–1226
Ben-Bassat A, Gottlieb D (1992) Optimal international reserves and sovereign risk. J Int Econ 33:345–362
Bilson JFO, Frenkel JA (1979) Dynamic adjustment and the demand for international reserves. NBER Working Paper Series n. 407
Calvo G, Leiderman L, Reinhart CM (1996) Inflows of capital to developing countries in the 1990’s. J Econ Perspect 10:123–139
Cheung Y-W, Ito H (2007) A cross-country empirical analysis of international reserves. Mimeo, University of California, Santa Cruz
Cifarelli G, Paladino G (2005) The international reserve glut: is it for real? Studi e Discussioni n. 142, Dipartimento di Scienze Economiche, Università di Firenze
Cifarelli G, Paladino G (2007) The buffer stock model redux? An analysis of the dynamics of foreign reserve accumulation. Working Paper n. 2, Dipartimento di Scienze Economiche, Università di Firenze
Dennis JG (2006) Cats in Rats. Estima, Evanston IL
Dooley MP, Folkerts-Landau D, Garber P (2003) An essay on the revised Bretton Woods System. NBER Working Paper Series n. 9971
Eichengreen B (2004) Global imbalances and the lesson of Bretton Woods. NBER Working Paper Series n. 10497
Escribano A, Granger CWJ (1998) Investigating the relationship between gold and silver prices. J Forecast 17:81–107
Frankel J (1999) No single currency regime is right for all countries or at all times. NBER Working Paper Series n. 7338
Frenkel JA, Jovanovic B (1981) Optimal international reserves: a stochastic framework. Econ J 91:507–514
García P, Soto C (2004) Large hoardings of international reserves: are they worth it? Working Paper n. 299, Central Bank of Chile
Gonzalo J, Ng S (2001) A systematic framework for analyzing the dynamic effects of permanent and transitory shocks. J Econ Dyn Control 25:1527–1546
Granger CWJ, Lee T-H (1989) Investigation of production, sales, and inventory relationships using multicointegration and non symmetric error correction models. J Appl Econ 4:145–159
Hansen H, Johansen S (1999) Some tests for parameter constancy in cointegrated VAR-models. Econom J 2:306–333
Heller RH (1966) Optimal international reserves. Econ J 76:296–311
Johansen S (1991) Estimation and hypothesis testing of cointegration vectors in Gaussian vector autoregressive models. Econometrica 59:1551–1580
Johansen S (1995) Likelihood based inference in cointegrated vector autoregressive models. Oxford University Press, Oxford
Lettau M, Ludvigson SC (2004) Understanding trend and cycle in asset values: reevaluating the wealth effect on consumption. Am Econ Rev 94:276–299
Kenen PB, Yudin EB (1965) The demand for international reserves. Rev Econ Stat 47:242–250
Maddala GS, Kim I-M (1998) Unit roots, cointegration, and structural change. Cambridge University Press, Cambridge
Ng S, Perron P (2001) Lag length selection and the construction of unit root tests with good size and power. Econometrica 69:1519–1554
Nielsen HB (2004) Cointegration analysis in the presence of outliers. Econom J 7:249–271
Perron P, Rodríguez G (2003) Searching for additive outliers in nonstationary time series. J Time Ser Anal 24:193–220
Rodrik D (2006) The social cost of foreign reserve. NBER Working Paper Series n. 11952
Stock JH, Watson MW (1993) A simple estimator of cointegrating vectors in higher order integrated systems. Econometrica 61:783–820
Summers LH (2006) Reflections on global account imbalances and emerging market reserve accumulation. L. K. Jha Memorial Lecture, Reserve Bank of India, Mumbai 24 March, www.President.Harvard.edu/speeches/2006/0324_rbi.html
Author information
Authors and Affiliations
Corresponding author
Additional information
The authors are grateful to the referee for extremely useful suggestions.
Rights and permissions
About this article
Cite this article
Cifarelli, G., Paladino, G. The Buffer Stock Model Redux? An Analysis of the Dynamics of Foreign Reserve Accumulation. Open Econ Rev 20, 525–543 (2009). https://doi.org/10.1007/s11079-007-9074-0
Received:
Accepted:
Published:
Issue Date:
DOI: https://doi.org/10.1007/s11079-007-9074-0