Abstract
Recent theoretical and empirical analyses of the relation between the current account and the government budget balance suggest that the “twin deficits” relation is subject to structural changes. Most previous empirical analyses impose the change point without resorting to econometric testing. In this paper we utilize time series data to evaluate the impact of structural breaks on the long- and short-run relation between current account, government balance and investment in 22 OECD countries. We found that when allowing for the possible existence of structural breaks of unknown date, the data reveal more clearly the long-run relation between the current account and its determinants. Moreover, the empirical results show that the degree of financial integration is generally increasing in most OECD countries, including the leading non-EU economies. This contrasts some recent evidence on the persistence of the so-called Feldstein–Horioka puzzle.
Similar content being viewed by others
Notes
According to the “current account targeting” hypothesis (Summers 1988) the government of a country may resort to fiscal policy to adjust its external position: this leads to a reverse causality nexus, running from the current account to the budget deficit.
This led some authors to investigate the so called “twin divergence” phenomenon (Kim and Roubini 2003).
Remark that \( ca_{t} = s_{t} - i_{t} \) (where the index t now refers to time), and since in the time series domain Eq. (3) becomes \( i_{t} = \alpha + \beta s_{t} + u_{t} \), it follows that \( ca_{t} = - \alpha + {\left( {1 - \beta } \right)}s_{t} - u_{t} \).
See for instance Gokhale et al. (2001) for savings dynamic in postwar United States.
Let \( \widehat{u}_{t} \) be the OLS residuals of Eq. (6). The auxiliary regression of the CRADF test is \( \Delta \widehat{u}_{t} = \gamma \widehat{u}_{{t - 1}} + {\sum\nolimits_{j = 1}^p {\Delta \widehat{u}_{{t - j}} + \eta _{t} } } \), where η t is a white noise. Starting from p = 5 we reduced the model until the Q statistics of the auxiliary regression residuals did not signal the presence of autocorrelation.
The order of lags in the auxiliary ADF regression was selected as explained in the preceding footnote 4.
We utilized as a trimming parameter 5, corresponding to about 11% of the available observations.
In the case of Canada the I(0) variable i t is omitted from the long-run regression.
The private savings ratio increased sharply by about seven GDP points from the end of the Sixties to the mid-Seventies.
The endogenously determined change point dates for the long-run relation fall respectively in 1980 and 1994, rather than in 1989.
The ADF* statistics for Canada however are unable to reject the null of non-cointegration.
The cointegration test in the 1975–2005 subsample was performed using the small sample critical values of Blangiewicz and Charemza (1990).
We excluded from the sample due to lack of data the following countries: Czech Republic, Hungary, Iceland, Korea, Luxembourg, Mexico, Poland and Slovak Republic.
References
Andrews DWK (1993) Tests for parameter instability and structural change with unknown change point. Econometrica 61:821–856
Bai J, Perron P (1998) Estimating and testing linear models with multiple structural changes. Econometrica 66:47–78
Blanchard O (1985) Debt, deficits, and finite horizons. J Polit Econ 93:223–247
Blanchard O, Giavazzi F (2002) Current account deficits in the Euro Area. The end of the Feldstein Horioka puzzle? Working paper No. 03-05. MIT Press, Massachusetts
Blangiewicz M, Charemza WW (1990) Cointegration in small samples: empirical percentiles, drifting moments and customized testing. Oxf Bull Econ Stat 52:303–315
Brown RL, Durbin J, Evans JM (1975) Techniques for testing the constancy of regression relationships over time. J R Stat Soc B 149–192
Chow GC (1960) Tests of equality between sets of coefficients in two linear regressions. Econometrica 28:591–605
Coakley J, Kulasi F, Smith R (2003) The Feldstein–Horioka puzzle and capital mobility: a review. In: Eijffinger SCW, Lemmen JJG (eds) International financial integration, vol 1. Elgar, Cheltenam (UK), Northampton (US)
Dickey D, Fuller W (1979) Distribution of the estimators for autoregressive time-series with a unit root. J Am Stat Assoc 74:427–431
Dickey D, Fuller W (1981) Likelihood ratio statistics for autoregressive time series with a unit root. Econometrica 49:1057–1072
Dolado JJ, Jenkinson T, Sosvilla-Rivero S (1990) Cointegration and unit roots. J Econ Surv 4:249–273
Enders W (2004) Applied econometric time-series, 2nd edn. Wiley, New York
Engle RF (1984) Wald, likelihood ratio, and Lagrange multiplier tests in econometrics, chap 13. In: Griliches Z, Intriligator MD (eds) Handbook of econometrics, vol II. Elsevier, Amsterdam
Engle RF, Granger CWJ (1987) Cointegration and error correction: representation, estimation and testing. Econometrica 55:251–276
Engsted TJ, Gonzalo J, Haldrup N (1997) Testing for multicointegration. Economics Letters 56:259–266
Feldstein M, Horioka C (1980) Domestic saving and international capital flows. Econ J 90:314–329
Fidrmuc J (2003) The Feldstein–Horioka puzzle and twin deficits in selected countries. Econ Plan 36:135–152
Gokhale J, Kotlikoff LJ, Sabelhaus J (2001) Understanding the postwar decline in U.S. saving: a cohort analysis. In: Kotlikoff L (ed) Essays on saving, altruism, and life-cycle planning. MIT Press, Cambridge, London
Granger CWJ (1969) Investigating causal relations by econometric models and cross-spectral methods. Econometrica 37:424–438
Gregory AW, Hansen BE (1996) Residual-based tests for cointegration in models with regime shifts. J Econ 70:99–126
Gundlach E, Sinn S (1992) Unit root tests of the current account balance: implications for international capital mobility. Appl Econ 24:617–625
Hatemi A, Shukur G (2002) Multivariate-based causality tests of twin deficits in the US. J Appl Stat 29:817–824
Johansen S (1988) Statistical analysis of cointegration vectors. J Econ Dyn Control 12:231–254
Kim S, Roubini N (2003) Twin deficit or twin divergence? Fiscal policy, current account, and real exchange rate in the US. Mimeo, http://www.stern.nyu.edu/globalmacro/
Kraay A, Ventura J (2002) Current accounts in the long and the short run. NBER Working Papers, No. 9030, July
Leachman LL, Francis B (2002) Twin deficits: apparition or reality? Appl Econ 34:1121–1132
Lemmen JJG, Eijffinger SCW (1995) The quantity approach to financial integration: the Feldstein–Horioka criterion revisited. Open Econ Rev 6:342–362
Maddala GS, Kim I-M (1998) Unit roots, cointegration and structural change. Cambridge University Press, Cambridge
Makin AJ (2004) The current account, fiscal policy, and medium-run income determination. Contemp Econ Policy 22:309–317
Mankiw NG (2002) Macroeconomics, fifth edn. Worth, New York
Mann CL (2002) Perspectives on the US current account deficit and sustainability. J Econ Perspect 16:131–152
Milesi-Ferretti GM, Lane P (1999) The external wealth of nations: measures of foreign assets and liabilities for industrial and developing countries. IMF Working Paper, No. 99/115
Normandin M (1999) Budget deficit persistence and the twin deficits hypothesis. J Int Econ 49:171–193
Obstfeld M, Rogoff K (1995) The intertemporal approach to the current account, chap 34. In: Grossman G, Rogoff K (eds) Handbook of international economics, vol III. Elsevier, Amsterdam
OECD (2004) OECD Statistical Compendium, CD-Rom edition
Roubini N, Setser B (2004) The US as a net debtor: the sustainability of the US external imbalances. Mimeo, November, http://www.stern.nyu.edu/globalmacro/
Summers LH (1988) Tax policy and international competitiveness. In: Frenkel J (ed) International aspects of fiscal policies. Chicago University Press, Chicago
Author information
Authors and Affiliations
Corresponding author
Additional information
I thank two anonymous referees for the many useful suggestions. Eleonora Pierucci supplied valuable research assistance.
Appendix: data sources
Appendix: data sources
Our data sample consists of annual data running from 1960 to 2005 (data for 2005 are the latest OECD projections) for 22 OECD countries.Footnote 15 The main data source is the 2004#2 CD-ROM edition of the “OECD Economic Outlook Statistics and Projections”. The Economic Outlook codes for the three variables in our dataset are as follows ca t : CBGDPR; \( s^{{\text{G}}}_{t} \): NLGQ; i t : 100 × ITV / GDPV.
In some cases missing values of the current account balance ca t were approximated by the trade balance, reconstructed as 100 × (XGSV − MGSV) / GDPV. This applies to Austria (1960–1969), Belgium (1960–1974), Denmark (1960–1974), Finland (1960–1974), France (1960–1974), Greece (1960–1974), Ireland (1960–1974), Italy (1960–1970), Japan (1960–1984), Netherlands (1960–1974), Norway (1960–1974), New Zealand (1961–1970), Portugal (1960–1974), Spain (1960–1974), Turkey (1967–1974), Sweden (1960–1974).
Supplementary data sources were utilized when the time series in OECD (2004) did not start in 1960.
The IFS 2004#9 CD-ROM edition was utilized for reconstructing the time series of the trade balance in: Switzerland (1960–1971), and the time series of \( s^{{\text{G}}}_{t} \) (defined as the ratio of the general government balance, IFS series code 80..ZF or 80G.ZF, to the GDP, IFS series code 99B.ZF...) in the following cases: Austria (1960–1963), Canada (1960–1980), Ireland (1960–1976), Netherlands (1960–1968), Norway (1960–1974), New Zealand (1960–1986), Switzerland (1960–1989), Turkey (1967–1978).
The 1999#1 edition of the OSC was utilized for reconstructing the time series of i t for Denmark (1960–1965).
Rights and permissions
About this article
Cite this article
Bagnai, A. Structural breaks and the twin deficits hypothesis. IEEP 3, 137–155 (2006). https://doi.org/10.1007/s10368-006-0050-8
Published:
Issue Date:
DOI: https://doi.org/10.1007/s10368-006-0050-8