EC729 F95 SQL/SUR

EC 729: Applied Econometrics I

Fall 1997

USING SQL AND SUR FOR
ESTIMATION OF POOLED CROSS-SECTION/TIME-SERIES MODELS

Step 1: We want to extract a number of variables for several countries from the IFS (IMF International Financial Statistics) database. This database contains a number of SQL tables:

COUNTRY: contains CCODE (int) and COUNTRY (char, name of country)

SERIES: contains SERIES (char, code given in hardcopy IFS) and NCODE (int)

ITEM: contains CCODE (int), NCODE (int), DESCRIPTION (char) etc.

ANNUAL: contains CCODE (int), NCODE (int), YEAR (int), Y_DATA (float)

There are also MONTHLY and QUARTERLY tables which we will not access.

The CCODE values can be found in IFS hardcopy values, or by using
$qbf -t ifs to search the COUNTRY table for desired countries.

The NCODE values are purely for internal use; if you consult the IFS hardcopy volumes you will find series identified by their SERIES code, or permutations thereof. For instance, Gross Domestic Product is SERIES 9C9B.

You can verify this by using qbf on the ITEM table.

Not all countries have all SERIES (or a similar series, in terms of definition, may be located under a slightly different SERIES code). Also, not all timeseries are of the same length. Thus after identifying several series that we might want to use:

6 1: Long term gov't bond yield

6 3: Wholesale price index

6 3A: Industrial price index

9C8C: Imports of goods and services

9C9B: G D P

and a set of candidate countries on which we will estimate a model of import demand, we use SQL to determine which of our countries have the appropriate coverage. The SQL select statement shown is placed in a file, IFS1.SQL, and executed via $SQLBATCH IFS1.SQL IFS1.OUT IFS. The results:

select country,series,count(year),min(year),max(year)

from annual,country,series

where country.ccode=annual.ccode

and annual.ncode=series.ncode

and country.ccode in (111,112,132,136,138,156,158)

and series in ('6 1','6 3','6 3A','9C8C','9C9B')

and year between 57 and 95

group by country,series

order by 1,2,3

+--------------------+------+-------------+------+------+

|country |series|col3 |col4 |col5 |

+--------------------+------+-------------+------+------+

|CANADA |6 1 | 39| 57| 95|

|CANADA |6 3 | 39| 57| 95|

|CANADA |9C8C | 39| 57| 95|

|CANADA |9C9B | 39| 57| 95|

|FRANCE |6 1 | 39| 57| 95|

|FRANCE |6 3A | 16| 80| 95|

|FRANCE |9C8C | 39| 57| 95|

|FRANCE |9C9B | 39| 57| 95|

|ITALY |6 1 | 39| 57| 95|

|ITALY |6 3 | 36| 57| 92|

|ITALY |6 3A | 4| 89| 92|

|ITALY |9C8C | 38| 57| 94|

|ITALY |9C9B | 38| 57| 94|

|JAPAN |6 1 | 30| 66| 95|

|JAPAN |6 3 | 39| 57| 95|

|JAPAN |9C8C | 39| 57| 95|

|JAPAN |9C9B | 39| 57| 95|

|NETHERLANDS |6 1 | 39| 57| 95|

|NETHERLANDS |6 3 | 39| 57| 95|

|NETHERLANDS |9C8C | 39| 57| 95|

|NETHERLANDS |9C9B | 39| 57| 95|

|UNITED KINGDOM |6 1 | 39| 57| 95|

|UNITED KINGDOM |6 3 | 39| 57| 95|

|UNITED KINGDOM |9C8C | 39| 57| 95|

|UNITED KINGDOM |9C9B | 39| 57| 95|

|UNITED STATES |6 1 | 39| 57| 95|

|UNITED STATES |6 3 | 39| 57| 95|

|UNITED STATES |6 3A | 39| 57| 95|

|UNITED STATES |9C8C | 39| 57| 95|

|UNITED STATES |9C9B | 39| 57| 95|

+--------------------+------+-------------+------+------+

(30 rows)

We see that some countries do not have certain series at all (e.g. Canada does not report a "6 3A") and other series are incomplete (e.g. France and Italy do not have all observations on that series). Thus we drop "6 3A" from consideration and decide to use Canada, Netherlands, UK and US in the analysis of the remaining four series. The retrieval of those series is performed by IFS2.SQL (output truncated):

select country,series,annual.ncode,year,y_data

from annual,country,series

where country.ccode=annual.ccode

and annual.ncode=series.ncode

and country.ccode in (111,112,138,156)

and series in ('6 1','6 3','9C8C','9C9B')

and year between 57 and 95

order by 1,2,3,4

+--------------------+------+------+------+-----------+

+--------------------+------+------+------+-----------+

|CANADA |6 1 | 309| 57| 4.110|

|CANADA |6 1 | 309| 58| 4.150|

|CANADA |6 1 | 309| 59| 5.070|

|CANADA |6 1 | 309| 60| 5.180|

|CANADA |6 1 | 309| 61| 5.050|

|CANADA |6 1 | 309| 62| 5.110|

|CANADA |6 1 | 309| 63| 5.090|

|CANADA |6 1 | 309| 64| 5.180|

|CANADA |6 1 | 309| 65| 5.210|

|CANADA |6 1 | 309| 66| 5.690|

|CANADA |6 1 | 309| 67| 5.940|

|CANADA |6 1 | 309| 68| 6.750|

|CANADA |6 1 | 309| 69| 7.580|

|CANADA |6 1 | 309| 70| 7.910|

|CANADA |6 1 | 309| 71| 6.950|

|CANADA |6 1 | 309| 72| 7.230|

|CANADA |6 1 | 309| 73| 7.560|

|CANADA |6 1 | 309| 74| 8.900|

|CANADA |6 1 | 309| 75| 9.030|

|CANADA |6 1 | 309| 76| 9.180|

|CANADA |6 1 | 309| 77| 8.700|

|CANADA |6 1 | 309| 78| 9.270|

|CANADA |6 1 | 309| 79| 10.210|

|CANADA |6 1 | 309| 80| 12.480|

|CANADA |6 1 | 309| 81| 15.220|

|CANADA |6 1 | 309| 82| 14.260|

|CANADA |6 1 | 309| 83| 11.790|

|CANADA |6 1 | 309| 84| 12.750|

|CANADA |6 1 | 309| 85| 11.040|

|CANADA |6 1 | 309| 86| 9.520|

|CANADA |6 1 | 309| 87| 9.950|

|CANADA |6 1 | 309| 88| 10.220|

|CANADA |6 1 | 309| 89| 9.920|

|CANADA |6 1 | 309| 90| 10.850|

|CANADA |6 1 | 309| 91| 9.760|

|CANADA |6 1 | 309| 92| 8.760|

|CANADA |6 1 | 309| 93| 7.840|

|CANADA |6 1 | 309| 94| 8.630|

|CANADA |6 1 | 309| 95| 8.280|

|CANADA |6 3 | 312| 57| 23.000|

|CANADA |6 3 | 312| 58| 23.000|

|CANADA |6 3 | 312| 59| 23.200|

|CANADA |6 3 | 312| 60| 23.300|

|CANADA |6 3 | 312| 61| 23.300|

|CANADA |6 3 | 312| 62| 23.600|

|CANADA |6 3 | 312| 63| 23.900|

|CANADA |6 3 | 312| 64| 24.100|

|CANADA |6 3 | 312| 65| 24.400|

|CANADA |6 3 | 312| 66| 25.100|

|CANADA |6 3 | 312| 67| 25.600|

|CANADA |6 3 | 312| 68| 26.100|

|CANADA |6 3 | 312| 69| 27.100|

etc.

This ordering gives us all variables for the first country, in a specific order, followed by all variables for the second country, etc. If we were estimating a model in which the data were "stacked" such as a fixed-effects model we would probably want to reverse the country and series sort order. For SUR, it doesn't matter. We convert the output file to get rid of headings, vertical bars, etc. via "$SQLC", using the VAX option (space delimited data) and excluding headers from IFS2.OUT, yielding the data file IFS2.DAT.

We will pull these data in to RATS, take logs of all but the interest rate variables, run OLS on each country and then run some variants of SUR (Zellner's Seemingly Unrelated Regressions estimator). To get the fastest turnaround on this job, I have placed the RATS commands in IFS2.RAT, and executed the job via $ ecb -d rats ifs2.rat/out=ifs2.lis, which runs a batch job on the Alpha AXP and returns the results as ifs2.lis:

VAX/VMS or AXP/OpenVMS RATS 4.20. Run on Jan 14 1997

* IFS2 Estimate pooled cross-section / time series models on IFS data

* CFB EC 729 Spr 97

compute ncty=4

compute nvar=4

compute nser=ncty*nvar

cal 1957 1 1

all nser 1995:1

open data ifs2.dat

do i=1,nser

(01.0025) data(org=obs,for='(46x,f13.0)') / [series]i

(01.0051) enddo

eqv 1 to nser

CDNbond CDNwp CDNimpt CDNgdp NLbond NLwp NLimpt NLgdp $

UKbond UKwp UKimpt UKgdp USbond USwp USimpt USgdp

tab

Series Obs Mean Std Error Minimum Maximum

CDNBOND 39 8.36717949 2.76510202 4.11000000 15.22000000

CDNWP 39 57.14871795 33.04050044 23.00000000 117.30000000

CDNIMPT 39 76.07846154 75.30976677 6.56000000 271.29000000

CDNGDP 39 290.17615385 252.87860652 34.47000000 776.30000000

NLBOND 39 7.14051282 1.90771970 3.91000000 11.55000000

NLWP 39 69.42564103 25.73552447 38.30000000 106.40000000

NLIMPT 39 131.08974359 96.14051886 16.10000000 298.90000000

NLGDP 39 267.86666667 194.36472678 34.70000000 634.70000000

UKBOND 39 9.43564103 2.91925801 5.19000000 14.77000000

UKWP 39 48.68717949 38.07791306 11.40000000 120.70000000

UKIMPT 39 59.08256410 59.99008075 4.58000000 203.09000000

UKGDP 39 224.92461538 220.58489694 21.92000000 700.89000000

USBOND 39 7.23000000 2.71199965 3.32000000 13.91000000

USWP 39 59.55384615 30.40637785 26.80000000 107.30000000

USIMPT 39 262.57692308 261.45753652 19.90000000 902.00000000

USGDP 39 2612.03589744 2133.35477834 448.60000000 7253.80000000

do i=2,14,4

(01.0028) log [series]i

(01.0040) log [series](i+1)

(01.0055) log [series](i+2)

(01.0070) enddo

tab

Series Obs Mean Std Error Minimum Maximum

CDNBOND 39 8.3671794872 2.7651020241 4.1100000000 15.2200000000

CDNWP 39 3.8668863718 0.6167029777 3.1354942159 4.7647347557

CDNIMPT 39 3.7056430997 1.2371923804 1.8809906030 5.6031883594

CDNGDP 39 5.1827671926 1.0793854806 3.5400893805 6.6545390434

NLBOND 39 7.1405128205 1.9077197035 3.9100000000 11.5500000000

NLWP 39 4.1687327388 0.3889551326 3.6454498962 4.6672055769

NLIMPT 39 4.4871625161 0.9919630543 2.7788192720 5.7001090693

NLGDP 39 5.2197400228 0.9681753276 3.5467396870 6.4531524463

UKBOND 39 9.4356410256 2.9192580121 5.1900000000 14.7700000000

UKWP 39 3.5268463105 0.8930658246 2.4336133554 4.7933081281

UKIMPT 39 3.3823528712 1.3138443547 1.5216989981 5.3136492305

UKGDP 39 4.8101702234 1.1962472355 3.0873994621 6.5523509560

USBOND 39 7.2300000000 2.7119996507 3.3200000000 13.9100000000

USWP 39 3.9466051072 0.5475512707 3.2884018875 4.6756286496

USIMPT 39 4.8946193827 1.3018749772 2.9907197317 6.8046145201

USGDP 39 7.4928457530 0.9171750083 6.1061316221 8.8892807485

compute eqn=0

do i=3,15,4

(01.0028) compute eqn=eqn+1

(01.0043) linreg(define=eqn) [series]i

(01.0066) # constant [series](i-2) [series](i-1) [series](i+1)

(01.0111) enddo

Dependent Variable CDNIMPT - Estimation by Least Squares

Annual Data From 1957:01 To 1995:01

Usable Observations 39 Degrees of Freedom 35

Centered R**2 0.996784 R Bar **2 0.996508

Uncentered R**2 0.999685 T x R**2 38.988

Mean of Dependent Variable 3.7056430997

Std Error of Dependent Variable 1.2371923804

Standard Error of Estimate 0.0731111299

Sum of Squared Residuals 0.1870833060

Regression F(3,35) 3615.5193

Significance Level of F 0.00000000

Durbin-Watson Statistic 0.651648

Q(9-0) 21.138496

Significance Level of Q 0.01204948

Variable Coeff Std Error T-Stat Signif

*******************************************************************************

1. Constant -2.274070824 0.125416697 -18.13212 0.00000000

2. CDNBOND -0.012145604 0.006883057 -1.76457 0.08636310

3. CDNWP 0.027018608 0.114790337 0.23537 0.81529061

4. CDNGDP 1.153218013 0.065551305 17.59260 0.00000000

Dependent Variable NLIMPT - Estimation by Least Squares

Annual Data From 1957:01 To 1995:01

Usable Observations 39 Degrees of Freedom 35

Centered R**2 0.998214 R Bar **2 0.998061

Uncentered R**2 0.999919 T x R**2 38.997

Mean of Dependent Variable 4.4871625161

Std Error of Dependent Variable 0.9919630543

Standard Error of Estimate 0.0436792017

Sum of Squared Residuals 0.0667755431

Regression F(3,35) 6521.2024

Significance Level of F 0.00000000

Durbin-Watson Statistic 1.177785

Q(9-0) 16.565831

Significance Level of Q 0.05596532

Variable Coeff Std Error T-Stat Signif

*******************************************************************************

1. Constant -2.277686420 0.208044115 -10.94809 0.00000000

2. NLBOND 0.029525609 0.005722244 5.15980 0.00000989

3. NLWP 0.676057964 0.096618644 6.99718 0.00000004

4. NLGDP 0.715690046 0.042501093 16.83933 0.00000000

Dependent Variable UKIMPT - Estimation by Least Squares

Annual Data From 1957:01 To 1995:01

Usable Observations 39 Degrees of Freedom 35

Centered R**2 0.996744 R Bar **2 0.996465

Uncentered R**2 0.999583 T x R**2 38.984

Mean of Dependent Variable 3.3823528712

Std Error of Dependent Variable 1.3138443547

Standard Error of Estimate 0.0781200007

Sum of Squared Residuals 0.2135957080

Regression F(3,35) 3571.1592

Significance Level of F 0.00000000

Durbin-Watson Statistic 0.852502

Q(9-0) 36.296742

Significance Level of Q 0.00003512

Variable Coeff Std Error T-Stat Signif

*******************************************************************************

1. Constant -1.962760762 0.055497414 -35.36671 0.00000000

2. UKBOND 0.023788196 0.005080053 4.68267 0.00004168

3. UKWP 0.107476383 0.147613926 0.72809 0.47140119

4. UKGDP 0.985745586 0.110977206 8.88241 0.00000000

Dependent Variable USIMPT - Estimation by Least Squares

Annual Data From 1957:01 To 1995:01

Usable Observations 39 Degrees of Freedom 35

Centered R**2 0.996439 R Bar **2 0.996134

Uncentered R**2 0.999770 T x R**2 38.991

Mean of Dependent Variable 4.8946193827

Std Error of Dependent Variable 1.3018749772

Standard Error of Estimate 0.0809464344

Sum of Squared Residuals 0.2293313835

Regression F(3,35) 3264.7977

Significance Level of F 0.00000000

Durbin-Watson Statistic 0.530783

Q(9-0) 49.620571

Significance Level of Q 0.00000013

Variable Coeff Std Error T-Stat Signif

*******************************************************************************

1. Constant -5.402807917 0.123856354 -43.62156 0.00000000

2. USBOND 0.020237804 0.007577641 2.67073 0.01140501

3. USWP 0.342000076 0.145021326 2.35827 0.02407581

4. USGDP 1.174636315 0.079669801 14.74381 0.00000000

sur(vcv) 4

# 1

# 2

# 3

# 4

Dependent Variable CDNIMPT - Estimation by Seemingly Unrelated Regressions

Annual Data From 1957:01 To 1995:01

Usable Observations 39 Degrees of Freedom 35

Centered R**2 0.996626 R Bar **2 0.996336

Uncentered R**2 0.999669 T x R**2 38.987

Mean of Dependent Variable 3.7056430997

Std Error of Dependent Variable 1.2371923804

Standard Error of Estimate 0.0748844637

Sum of Squared Residuals 0.1962689018

Durbin-Watson Statistic 0.586200

Variable Coeff Std Error T-Stat Signif

*******************************************************************************

1. Constant -2.383020624 0.114222738 -20.86293 0.00000000

2. CDNBOND -0.008546393 0.006268742 -1.36333 0.17277706

3. CDNWP 0.156973311 0.103267909 1.52006 0.12849615

4. CDNGDP 1.071469113 0.059315640 18.06385 0.00000000

Dependent Variable NLIMPT - Estimation by Seemingly Unrelated Regressions

Annual Data From 1957:01 To 1995:01

Usable Observations 39 Degrees of Freedom 35

Centered R**2 0.998205 R Bar **2 0.998051

Uncentered R**2 0.999918 T x R**2 38.997

Mean of Dependent Variable 4.4871625161

Std Error of Dependent Variable 0.9919630543

Standard Error of Estimate 0.0437965271

Sum of Squared Residuals 0.0671347524

Durbin-Watson Statistic 1.161229

Variable Coeff Std Error T-Stat Signif

*******************************************************************************

5. Constant -2.253428515 0.194493583 -11.58613 0.00000000

6. NLBOND 0.031404700 0.005304105 5.92083 0.00000000

7. NLWP 0.664282887 0.090243844 7.36098 0.00000000

8. NLGDP 0.717876280 0.039570930 18.14151 0.00000000

Dependent Variable UKIMPT - Estimation by Seemingly Unrelated Regressions

Annual Data From 1957:01 To 1995:01

Usable Observations 39 Degrees of Freedom 35

Centered R**2 0.996460 R Bar **2 0.996157

Uncentered R**2 0.999546 T x R**2 38.982

Mean of Dependent Variable 3.3823528712

Std Error of Dependent Variable 1.3138443547

Standard Error of Estimate 0.0814465527

Sum of Squared Residuals 0.2321739331

Durbin-Watson Statistic 0.806174

Variable Coeff Std Error T-Stat Signif

*******************************************************************************

9. Constant -1.929180157 0.051939217 -37.14304 0.00000000

10. UKBOND 0.015286051 0.004244541 3.60134 0.00031658

11. UKWP 0.156281959 0.133000724 1.17505 0.23997634

12. UKGDP 0.959657699 0.100146550 9.58253 0.00000000

Dependent Variable USIMPT - Estimation by Seemingly Unrelated Regressions

Annual Data From 1957:01 To 1995:01

Usable Observations 39 Degrees of Freedom 35

Centered R**2 0.996357 R Bar **2 0.996045

Uncentered R**2 0.999765 T x R**2 38.991

Mean of Dependent Variable 4.8946193827

Std Error of Dependent Variable 1.3018749772

Standard Error of Estimate 0.0818709481

Sum of Squared Residuals 0.2345998252

Durbin-Watson Statistic 0.548352

Variable Coeff Std Error T-Stat Signif

*******************************************************************************

13. Constant -5.339794014 0.115438846 -46.25647 0.00000000

14. USBOND 0.025281013 0.006226198 4.06043 0.00004898

15. USWP 0.365815148 0.126355728 2.89512 0.00379012

16. USGDP 1.148816354 0.070389297 16.32090 0.00000000

Covariance\Correlation Matrix of Coefficients

Constant CDNBOND CDNWP CDNGDP

Constant 0.01304683391 0.1814286576 -0.8693579264 0.7302552634

CDNBOND 0.00012990888 0.00003929712 -0.0834731272 -0.1296016697

CDNWP -0.01025454907 -0.00005403715 0.01066426100 -0.9617049408

CDNGDP 0.00494762171 -0.00004819036 -0.00589082951 0.00351834519

Constant -0.00162632144 0.00004654802 0.00149309546 -0.00087670100

NLBOND -0.00001023379 -0.00000336998 0.00000276938 0.00000534891

NLWP 0.00070194829 -0.00002138741 -0.00072854971 0.00044266372

NLGDP -0.00023637027 0.00001277339 0.00029201815 -0.00019289078

Constant 0.00079531568 -0.00001218804 0.00006100259 -0.00017121412

UKBOND 0.00000796796 0.00000556185 0.00000150385 -0.00001163861

UKWP -0.00297804225 0.00006953948 0.00335243489 -0.00203892852

UKGDP 0.00201125151 -0.00005936311 -0.00247365777 0.00155337941

Constant 0.00174936857 0.00014143530 0.00151261249 -0.00168049387

USBOND 0.00015578663 0.00001629765 -0.00013531478 0.00004458909

USWP -0.00566175151 0.00000392897 0.00590119022 -0.00331682946

USGDP 0.00260799000 -0.00003667144 -0.00317956019 0.00192828350

Constant NLBOND NLWP NLGDP

Constant -0.0732063144 -0.0168916400 0.0680980955 -0.0522954644

CDNBOND 0.0381782111 -0.1013526969 -0.0378059599 0.0514931737

CDNWP 0.0743390345 0.0050559725 -0.0781764998 0.0714608541

CDNGDP -0.0759935969 0.0170013860 0.0826964971 -0.0821799718

Constant 0.03782775379 -0.3898394723 -0.9819559357 0.9194442661

NLBOND -0.00040216399 0.00002813353 0.3985391462 -0.5421650037

NLWP -0.01723514193 0.00019076587 0.00814395143 -0.9698022568

NLGDP 0.00707631146 -0.00011379412 -0.00346319569 0.00156585848

Constant 0.00037295706 -0.00000670187 0.00003199696 -0.00008435414

UKBOND 0.00002798049 0.00000386305 -0.00001620556 0.00000229742

UKWP -0.00491074970 0.00004388060 0.00237416091 -0.00101534226

UKGDP 0.00347195008 -0.00003835803 -0.00171561243 0.00075748516

Constant 0.00030100535 -0.00001646208 -0.00020751972 0.00013002379

USBOND 0.00003093493 -0.00000228485 -0.00001387080 0.00000827699

USWP 0.00062040994 0.00000493783 -0.00030963401 0.00012167519

USGDP -0.00039719543 0.00000180090 0.00020416926 -0.00008942802

Constant UKBOND UKWP UKGDP

Constant 0.1340576431 0.0164347864 -0.1960307862 0.1758238649

CDNBOND -0.0374333029 0.2090299422 0.0834059504 -0.0945584369

CDNWP 0.0113733264 0.0034308916 0.2440849281 -0.2391873679

CDNGDP -0.0555744221 -0.0462275717 -0.2584513294 0.2615003816

Constant 0.0369197001 0.0338937259 -0.1898403458 0.1782510934

NLBOND -0.0243269810 0.1715883980 0.0622023057 -0.0722118111

NLWP 0.0068264621 -0.0423073244 0.1978055883 -0.1898303300

NLGDP -0.0410425912 0.0136783554 -0.1929221980 0.1911445345

Constant 0.00269768224 -0.2660094857 0.0832378279 -0.1611436089

UKBOND -0.00005864396 0.00001801613 0.1308394518 -0.1818619891

UKWP 0.00057500304 0.00007386241 0.01768919266 -0.9935950275

UKGDP -0.00083819386 -0.00007730520 -0.01323425225 0.01002933148

Constant 0.00275196429 0.00007778385 0.00358586454 -0.00333633796

USBOND -0.00003334777 0.00000955461 -0.00007829997 0.00004560044

USWP 0.00016581109 0.00000766909 0.00858764473 -0.00634602843

USGDP -0.00041117974 -0.00002363996 -0.00492627227 0.00374382657

Constant USBOND USWP USGDP

Constant 0.1326712368 0.2190557521 -0.3922864450 0.3243745181

CDNBOND 0.1954454014 0.4175627439 0.0049602477 -0.0831076349

CDNWP 0.1268850117 -0.2104538755 0.4522507495 -0.4374163808

CDNGDP -0.2454232667 0.1207359134 -0.4425465740 0.4618437133

Constant 0.0134065480 0.0255458790 0.0252451838 -0.0290129804

NLBOND -0.0268856515 -0.0691867151 0.0073676482 0.0048236039

NLWP -0.0199200219 -0.0246865747 -0.0271541451 0.0321414994

NLGDP 0.0284639121 0.0335948756 0.0243349721 -0.0321063341

Constant 0.4589817554 -0.1031213256 0.0252652287 -0.1124681854

UKBOND 0.1587474298 0.3615423814 0.0142994134 -0.0791241924

UKWP 0.2335543159 -0.0945550342 0.5110049741 -0.5262084052

UKGDP -0.2885905245 0.0731324455 -0.5015001705 0.5310960904

Constant 0.01332612713 0.1885850589 0.2877310155 -0.5045476246

USBOND 0.00013554459 0.00003876554 -0.4632735682 0.3114012223

USWP 0.00419694802 -0.00036446464 0.01596577011 -0.9689446502

USGDP -0.00409978202 0.00013647399 -0.00861788173 0.00495465306

Covariance\Correlation Matrix of Residuals

CDNIMPT NLIMPT UKIMPT USIMPT

CDNIMPT 0.00503253594 -0.1615043323 0.4179519725 0.5906776360

NLIMPT -0.00047535603 0.00172140391 0.2461139469 -0.0982453561

UKIMPT 0.00228767459 0.00078786543 0.00595317777 0.7316601238

USIMPT 0.00324994124 -0.00031614401 0.00437839915 0.00601538013

sur(iter=10) 4

# 1

# 2

# 3

# 4

SUR - Iterations Taken 7

Dependent Variable CDNIMPT - Estimation by Seemingly Unrelated Regressions

Annual Data From 1957:01 To 1995:01

Usable Observations 39 Degrees of Freedom 35

Centered R**2 0.996448 R Bar **2 0.996143

Uncentered R**2 0.999652 T x R**2 38.986

Mean of Dependent Variable 3.7056430997

Std Error of Dependent Variable 1.2371923804

Standard Error of Estimate 0.0768306427

Sum of Squared Residuals 0.2066031680

Durbin-Watson Statistic 0.547273

Variable Coeff Std Error T-Stat Signif

*******************************************************************************

1. Constant -2.435649920 0.115349137 -21.11546 0.00000000

2. CDNBOND -0.007280146 0.006148398 -1.18407 0.23638452

3. CDNWP 0.219585938 0.103117812 2.12947 0.03321567

4. CDNGDP 1.032863958 0.059685020 17.30525 0.00000000

Dependent Variable NLIMPT - Estimation by Seemingly Unrelated Regressions

Annual Data From 1957:01 To 1995:01

Usable Observations 39 Degrees of Freedom 35

Centered R**2 0.998153 R Bar **2 0.997994

Uncentered R**2 0.999916 T x R**2 38.997

Mean of Dependent Variable 4.4871625161

Std Error of Dependent Variable 0.9919630543

Standard Error of Estimate 0.0444263416

Sum of Squared Residuals 0.0690794940

Durbin-Watson Statistic 1.127531

Variable Coeff Std Error T-Stat Signif

*******************************************************************************

5. Constant -2.085760180 0.192279876 -10.84752 0.00000000

6. NLBOND 0.029910685 0.005174267 5.78066 0.00000001

7. NLWP 0.582976493 0.089007343 6.54976 0.00000000

8. NLGDP 0.752733250 0.038839677 19.38052 0.00000000

Dependent Variable UKIMPT - Estimation by Seemingly Unrelated Regressions

Annual Data From 1957:01 To 1995:01

Usable Observations 39 Degrees of Freedom 35

Centered R**2 0.996188 R Bar **2 0.995861

Uncentered R**2 0.999511 T x R**2 38.981

Mean of Dependent Variable 3.3823528712

Std Error of Dependent Variable 1.3138443547

Standard Error of Estimate 0.0845250668

Sum of Squared Residuals 0.2500570419

Durbin-Watson Statistic 0.762277

Variable Coeff Std Error T-Stat Signif

*******************************************************************************

9. Constant -1.912174145 0.055442659 -34.48922 0.00000000

10. UKBOND 0.012056456 0.003851986 3.12993 0.00174847

11. UKWP 0.191381314 0.133863328 1.42968 0.15280976

12. UKGDP 0.936722390 0.101044436 9.27040 0.00000000

Dependent Variable USIMPT - Estimation by Seemingly Unrelated Regressions

Annual Data From 1957:01 To 1995:01

Usable Observations 39 Degrees of Freedom 35

Centered R**2 0.996297 R Bar **2 0.995980

Uncentered R**2 0.999761 T x R**2 38.991

Mean of Dependent Variable 4.8946193827

Std Error of Dependent Variable 1.3018749772

Standard Error of Estimate 0.0825421956

Sum of Squared Residuals 0.2384624918

Durbin-Watson Statistic 0.538415

Variable Coeff Std Error T-Stat Signif

*******************************************************************************

13. Constant -5.309150832 0.115294330 -46.04867 0.00000000

14. USBOND 0.024752416 0.005074317 4.87798 0.00000107

15. USWP 0.421859469 0.116436550 3.62308 0.00029111

16. USGDP 1.115717283 0.066072665 16.88622 0.00000000

Covariance\Correlation Matrix of Residuals

CDNIMPT NLIMPT UKIMPT USIMPT

CDNIMPT 0.00529751713 -0.2340242035 0.4493338294 0.6240630247

NLIMPT -0.00071686806 0.00177126908 0.1911930250 -0.1798888884

UKIMPT 0.00261874097 0.00064431992 0.00641171902 0.7543533082

USIMPT 0.00355174986 -0.00059200359 0.00472323741 0.00611442287

compute unrest=%logdet

sur(iter=10) 4 / equate 4

# 1

# 2

# 3

# 4

SUR - Iterations Taken 11

Dependent Variable CDNIMPT - Estimation by Seemingly Unrelated Regressions

Annual Data From 1957:01 To 1995:01

Usable Observations 39 Degrees of Freedom 35

Centered R**2 0.995249 R Bar **2 0.994842

Uncentered R**2 0.999535 T x R**2 38.982

Mean of Dependent Variable 3.7056430997

Std Error of Dependent Variable 1.2371923804

Standard Error of Estimate 0.0888561649

Sum of Squared Residuals 0.2763396314

Durbin-Watson Statistic 0.396031

Variable Coeff Std Error T-Stat Signif

*******************************************************************************

2. Constant -2.598789726 0.096727206 -26.86721 0.00000000

3. CDNBOND -0.000379650 0.006099533 -0.06224 0.95036977

4. CDNWP 0.425243482 0.051803162 8.20883 0.00000000

1. CDNGDP 0.899758956 0.024101530 37.33203 0.00000000

Dependent Variable NLIMPT - Estimation by Seemingly Unrelated Regressions

Annual Data From 1957:01 To 1995:01

Usable Observations 39 Degrees of Freedom 35

Centered R**2 0.997143 R Bar **2 0.996898

Uncentered R**2 0.999870 T x R**2 38.995

Mean of Dependent Variable 4.4871625161

Std Error of Dependent Variable 0.9919630543

Standard Error of Estimate 0.0552448901

Sum of Squared Residuals 0.1068199258

Durbin-Watson Statistic 0.785383

Variable Coeff Std Error T-Stat Signif

*******************************************************************************

5. Constant -1.413224012 0.147365032 -9.58995 0.00000000

6. NLBOND 0.022865804 0.005189809 4.40590 0.00001053

7. NLWP 0.249621453 0.059289577 4.21021 0.00002551

1. NLGDP 0.899758956 0.024101530 37.33203 0.00000000

Dependent Variable UKIMPT - Estimation by Seemingly Unrelated Regressions

Annual Data From 1957:01 To 1995:01

Usable Observations 39 Degrees of Freedom 35

Centered R**2 0.996496 R Bar **2 0.996195

Uncentered R**2 0.999551 T x R**2 38.982

Mean of Dependent Variable 3.3823528712

Std Error of Dependent Variable 1.3138443547

Standard Error of Estimate 0.0810411998

Sum of Squared Residuals 0.2298686621

Durbin-Watson Statistic 0.807168

Variable Coeff Std Error T-Stat Signif

*******************************************************************************

8. Constant -1.923055220 0.052801217 -36.42066 0.00000000

9. UKBOND 0.017191887 0.003677483 4.67491 0.00000294

10. UKWP 0.231140742 0.035172817 6.57157 0.00000000

1. UKGDP 0.899758956 0.024101530 37.33203 0.00000000

Dependent Variable USIMPT - Estimation by Seemingly Unrelated Regressions

Annual Data From 1957:01 To 1995:01

Usable Observations 39 Degrees of Freedom 35

Centered R**2 0.994829 R Bar **2 0.994385

Uncentered R**2 0.999667 T x R**2 38.987

Mean of Dependent Variable 4.8946193827

Std Error of Dependent Variable 1.3018749772

Standard Error of Estimate 0.0975515881

Sum of Squared Residuals 0.3330709317

Durbin-Watson Statistic 0.421823

Variable Coeff Std Error T-Stat Signif

*******************************************************************************

11. Constant -5.096306760 0.116161070 -43.87276 0.00000000

12. USBOND 0.024952396 0.005256986 4.74652 0.00000207

13. USWP 0.777570890 0.054395122 14.29486 0.00000000

1. USGDP 0.899758956 0.024101530 37.33203 0.00000000

Covariance\Correlation Matrix of Residuals

CDNIMPT NLIMPT UKIMPT USIMPT

CDNIMPT 0.00708563157 -0.5361444625 0.4764024258 0.7681733680

NLIMPT -0.00236191868 0.00273897246 -0.0041177344 -0.5803854073

UKIMPT 0.00307872450 -0.00001654473 0.00589406826 0.6661357454

USIMPT 0.00597564229 -0.00280702564 0.00472613782 0.00854028030

compute rest=%logdet

compute lrt=-2.0*(unrest-rest)

* difference in the log det is distrib chi-squared with (nvar-1) d.f.

write unrest rest lrt

-23.4149

-23.1735

0.4828

cdf chisqr lrt nvar-1

Chi-Squared(3)= 0.482795 with Significance Level 0.92265465

end

Normal Completion

Restricting the "income" elasticity of import demand to be equal across countries cannot be rejected by the data.