Table 1.

Structural Changes of the Old-Age Survival Index

	Sweden	United Kingdom	Italy	United States	Japan	Taiwan	India
	1751−2002	1841−1998	1872−2000	1900−2001	1891−2001	1906−2002	1881−1997
t1	1876	1900			1947	1940	1951
t2					1989	1983	1989
t	0.0006 ***	0.0009 ***	0.0029 ***	0.0033 ***	0.0010 ***	0.0004	0.0392
	(0.0001)	(0.0001)	(0.0001)	(0.0002)	(0.0002)	(0.0010)	(0.0232)
(t-t1)d1	0.002 ***	0.0024 ***			0.0065 ***	0.0056 ***	0.3457 ***
	(0.0001)	(0.0001)			(0.0004)	(0.0011)	(0.0599)
(t-t2)d2					−0.0014 ***	−0.0024 ***	−0.3075 *
					(0.0007)	(0.0003)	(0.1568)
Dm 1773	−0.0231 ***	0.0004 ***	−0.1231 ***
	(0.0027)		(0.0023)
Dm 1808	−0.1625 ***
	(0.0063)
Dm 1918	−0.096 ***
	(0.0040)
Constant	0.2869 ***	0.3205 ***	0.2518 ***	0.2245 ***	0.3250 ***	0.2860 ***	22.7955 ***
	(0.0080)	(0.0049)	(0.0053)	(0.0195)	(0.0104)	(0.0322)	(0.8454)
Adjusted
R2	0.9251	0.9587	0.9471	0.9585	0.9872	0.9944	0.9628
N	252	158	129	48	59	43	20
P-value	0.0000	0.0000			0.0000	0.0000	0.0000
D.W.	0.5765	0.7026	0.397	0.0958	0.3897	0.3651	0.0338

Note. Dependent variables are adult survival index except for India. The dependent variable for India is life expectancy at age 30. The equation $q_t={\beta }_0+{\beta }_{1}t+{\beta }_2{d}_1(t-t_1)+{\beta }_3{d}_2(t-t_2)+{\epsilon }_t$ is estimated, where d₁ = 1 if t ≥ t₁ and 0 otherwise, and d₂=1 if t ≥ t₂ and 0 otherwise. Dm1773 is a dummy variable equal to 1 for year 1773. Dm1808, Dm1918 are defined in similar fashion N is number of observations. “P-value” is the p-value of F-test for the null hypothesis that β₁, β₂, and β₃ are all zero. D.W. is Durbin-Watson statistics. The low Durbin-Watson statistics imply serial correlation. The augmented Dickey-Fuller test indicates that some variables are not trend stationary.

^***denotes significant at 1 % level

^**denotes significant at 5 % level, and

^*denotes significant at 10 % level.