Private saving determinants in European countries: A panel cointegration approach

Abstract

This paper investigates the determinants of aggregate private saving in European countries employing panel data. The long-run saving function is estimated based on an extended lifecycle hypothesis taking into account the economic and demographic developments during this period. A long-run saving function sensitive to dependency ratio, old dependency ratio, liquidity, public finances, real disposable income growth, real interest rate and inflation is found to exist. The empirical evidence suggests the existence of a long-run saving function in Europe. The policy implications of such a relationship are presented.

Acknowledgments

The views expressed in this paper are those of the author and not of the Bank of Greece. The author wishes to thank two anonymous referees for their helpful comments in a previous version of this paper; Chiang and Kao who offer the NPT 1.3 as public program and Peter Pedroni for providing his program for the estimation of FMOLS. The author also wishes to acknowledge useful comments and suggestions by Stephen Hall. All errors and deficiencies are the responsibility of the author.

Notes

1 For an extensive discussion over savings in Europe, see, among others, Khaled and Thirlwall (1999) and Börsch-Supan and Brugiavini (2001). Data, not presented in Table 1 due to space limitations, show that savings ratio rises at a decreasing rate as per capita income increases and levels off at around 20–25% of national income.

2 Intertemporal substitution is the process of maximizing consumer's utility by allocating resources across time.

3 Total households’ loans, that is consumption and housing loans, as percentage of GDP in the Euro area were increased from 30.5% in 1997 to 36.8% in 2001 and 38% in 2002.

4 Specific country characteristics, which do not alter over time, such as religion, education, etc. might affect savings.

5 Panel cointegration modeling is applied to estimate the long-run relationship between savings and the other macroeconomic variables. For more details see Section 3.

6 Many empirical studies report that the response of savings to interest rate is not different to zero for developing countries where the population may not be able to borrow even at black market rates. In addition, this result may imply that interest rate is an effective policy tool, only in conjunction with financial markets.

7 The main deregulation and liberalization of financial system occurred in France in the second half of the 1980s to improve the efficiency of the allocation of financial resources. The creation of new financial instruments provided the private sector access to a broader range of financing, while the unification of the credit market favored the expansion of the French financial system.

8 Luxembourg and Portugal are excluded since there are no available data for the total period.

9 This version of the test is an extension of the Dickey-Fuller test, which makes a semiparametric correction for autocorrelation and is more robust in the case of weakly autocorrelated and heteroskedastic regression residuals.

10 The KPSS procedure assumes the univariate series can be decomposed into the sum of a deterministic trend, random walk and stationary I(0) disturbance and is based on a Lagrange Multiplier Score Testing Principle. This test reverses the null and the alternative hypothesis. A finding favorable to a unit root in this case requires strong evidence against the null hypothesis of stationarity.

11 The following equation is estimated: u_it = ρu_i(t − 1) + e_it where u_it are the estimated residuals.

12 When the regression analysis reveals the existence of a relationship, otherwise not expected, then such regression is called spurious.

13 A time series is stationary if its mean and variance do not vary systematically over time.

14 The unit root tests are not reported but are available from the author upon request.

15 In some countries there is more than one cointegrating vectors. From theory it seems likely that there will be additional long-run relationships between the demographic and economic variables. In addition, the Larsson, Lyhagen, and Lothgren (2001) test for panel cointegration is employed. The test is obtained by estimating the average of the N individual trace statistics and then standardizing. The test follows standard normal distribution. The test indicates the existence of three cointegrating vectors. However the other long-run relationships for the individual countries and for the panel are not examined since the purpose of the study is to estimate a private saving function.

16 The Pedroni (1999) panel cointegration statistics are calculated using Chiang and Kao (2002) NPT 1.3 program.

17 The FMOLS is estimated using a routine kindly provided by Pedroni.

18 Although, it may be concluded that inflation stabilization could have an adverse effect on saving, it should be emphasized that price stabilization affects saving through indirect channels that are likely to more than compensate for any negative direct effect of inflation. In this context, lower inflation raises growth and this might in turn increase saving. In addition, the macroeconomic stability implied from a fiscal discipline policy will have a positive effect on national saving (Loayza et al., 2000).

19 Private savings as percentage of GDP, since the currency changeover to Euro, remained almost constant. In particular, provisional figures indicate that the private savings to GDP ratio for the Euro area (excluding Luxembourg) increased to 20.8% in 2004 compared to 20.1% in 2002 and 19.6 in 2001.