Harvested Area, Climate Variability, and Cereal Production Dynamics in Algeria: Var Evidence, 1990–2024

Symbol	Variable and definition	Transformation	Unit	Source
LPROD	FAOSTAT “Cereals, primary” – Production; annual national cereal production.	Natural logarithm	Metric tonnes	FAOSTAT
LHA	FAOSTAT “Cereals, primary” – Area harvested; annual area that reaches harvest.	Natural logarithm	Hectares	FAOSTAT
LPREC	CCKP national (ADM0) precipitation indicator (pr); annual accumulated precipitation.	Natural logarithm	Millimetres	World Bank CCKP
LTEMP	CCKP national (ADM0) average surface-air temperature indicator (tas).	Natural logarithm; level-form robustness check	Degrees Celsius	World Bank CCKP
D1	Low-precipitation anomaly indicator: 1 if PREC < 67.72 mm; 0 otherwise.	Deterministic binary control	Binary	Authors’ calculation
D2	High-precipitation anomaly indicator: 1 if PREC > 95.83 mm; 0 otherwise.	Deterministic binary control	Binary	Authors’ calculation

[i] Source: own elaboration based on FAOSTAT and World Bank Climate Change Knowledge Portal data.

Table 2.

Descriptive statistics

Statistic	LPROD	LHA	LPREC	LTEMP
Mean	14.9322	14.7335	4.3888	3.1617
Median	15.0761	14.8036	4.4343	3.1633
Maximum	15.6182	15.1140	4.6774	3.1855
Minimum	13.6762	13.8720	4.0486	3.1280
Std. Dev.	0.5174	0.3185	0.1793	0.0125
Skewness	–1.0809	–1.2807	–0.4489	–0.8860
Kurtosis	3.2517	4.0177	1.9246	4.1667
Jarque-Bera	6.9089	11.0788	2.8618	6.5651
Probability (Jarque-Bera)	0.0316	0.0039	0.2390	0.0375
Observations	35	35	35	35

[i] Note: Statistics refer to the logged variables over 1990–2024.

[ii] Source: own elaboration.

Time-series plots of the log-transformed variables, 1990–2024
Source: own elaboration based on FAOSTAT and World Bank Climate Change Knowledge Portal data.

Table 3.

Unit-root and stationarity tests

Panel A. Conventional ADF and PP unit-root tests
Variable	ADF: C	ADF: C+T	ADF: none	PP: C	PP: C+T	PP: none	Decision
LPROD	–4.4949***	–5.7846***	0.1515	–4.6596***	–5.7846***	0.5140	I(0)
LHA	–5.7619***	–3.9701**	0.4212	–5.9625***	–6.3377***	0.0718	I(0)
LPREC	–3.3528**	–3.8020**	–0.6048	–3.2801**	–3.7718**	–1.4537	I(0)
LTEMP	–2.1619	–4.1701**	0.4554	–1.9491	–4.2693***	1.4484	Stationary under trend specification
Panel B. Zivot-Andrews unit-root test with one endogenous structural break
Variable	ZA statistic	Selected lag	Break year	5% critical value		Decision
LPROD	–7.1064	0	2002	–5.0733		Reject unit root
LHA	–4.6927	3	2016	–5.0733		Do not reject
LPREC	–5.6676	0	2002	–5.0733		Reject unit root
LTEMP	–5.8781	0	1996	–5.0733		Reject unit root

[i] Panel A: C = constant; C+T = constant and deterministic trend. The null hypothesis is a unit root. LTEMP rejects the unit-root null only when the trend is included.

[ii] Panel B: Model C allows one endogenous break in intercept and trend; maximum lag = 4; lag selected by AIC. Critical values: 1% = −5.5756, 5% = −5.0733, 10% = −4.8267.

[iii] Source: own elaboration.

VAR Stability Diagnostic Based on Inverse AR Roots
Source: own elaboration based on EViews stability diagnostics.

Table 4.

VAR lag order selection

Lag	LogL	LR	FPE	AIC	SC	HQ
0	142.6939	NA	3.34e-09	–8.168369	–7.618718	–7.986175
1	173.4450	48.04858^*	1.36e-09^*	–9.090312^*	–7.807793^*	–8.665193^*
2	181.3963	10.43609	2.46e-09	–8.587268	–6.571881	–7.919225
3	196.3854	15.92591	3.21e-09	–8.524087	–5.775832	–7.613118

* indicates the lag selected by the criterion.

Source: own elaboration.

Table 5.

Quandt-Andrews unknown-breakpoint tests for VAR parameter stability

VAR equation	Maximum LR F	Candidate year	p-value	Decision at 5%
LPROD	2.4787	2001	0.3049	Do not reject stability
LHA	4.1509	1997	0.0192	Reject stability
LPREC	2.1000	2003	0.4865	Do not reject stability
LTEMP	3.1885	1997	0.1052	Do not reject stability

[i] Note: The test searches the 15% trimmed sample (1997–2019). Candidate years correspond to the maximum LR F-statistic; a candidate year is not treated as a confirmed break when the stability null is not rejected.

[ii] Source: own elaboration.

Table 6.

VAR residual diagnostic tests

Diagnostic	Statistic	df	p-value	Additional statistic	Additional p	Decision
LM, lag 1	LRE = 7.6842	16	0.9577	Rao F = 0.4564	0.9584	No serial correlation
LM, lag 2	LRE = 16.8861	16	0.3930	Rao F = 1.0745	0.3978	No serial correlation
LM, lags 1–2	LRE = 32.4358	32	0.4453	Rao F = 1.0147	0.4686	No serial correlation
Joint skewness	–	4	0.3859	–	–	Normality not rejected
Joint kurtosis	–	4	0.9030	–	–	Normality not rejected
Joint Jarque-Bera	JB = 5.1957	8	0.7365	–	–	Normality not rejected
White heteroskedasticity	–	100	0.1359	–	–	Homoskedasticity not rejected

[i] Note: The respective null hypotheses are no residual serial correlation, multivariate normal residuals and homoskedasticity.

[ii] Source: own elaboration.

Table 7.

VAR Granger causality/block-exogeneity Wald tests for LPROD

Excluded variable	df	Chi-square	p-value
LHA	1	1.3587	0.2438
LPREC	1	0.1166	0.7327
LTEMP	1	2.9804	0.0843
All three	3	8.6413	0.0345

[i] Note: Included observations = 34 after lag adjustment. The test is interpreted as predictive content, not structural causality.

[ii] Source: own elaboration.

Bootstrap Impulse Response Functions of LPROD
Note: Shaded bands are 95% standard percentile-bootstrap confidence intervals based on 2,000 EViews repetitions. Benchmark Cholesky ordering: LPROD, LHA, LPREC, LTEMP.

Table 8.

Variance decomposition of total cereal production (LPROD)

Period	S.E.	LPROD (%)	LHA (%)	LPREC (%)	LTEMP (%)
1	0.4647	100.00	0.00	0.00	0.00
2	0.4827	93.07	2.92	0.12	3.87
4	0.5078	84.35	8.80	1.29	5.55
6	0.5194	80.82	11.22	1.69	6.24
8	0.5249	79.22	12.32	1.88	6.56
10	0.5276	78.47	12.85	1.97	6.71

[i] Values are descriptive point estimates of LPROD forecast-error variance shares under the benchmark Cholesky ordering. They are conditional on the model, horizon and identification scheme; uncertainty intervals are not reported, so no inferential significance is attached to differences among the shares.

[ii] Source: own elaboration.

Table 9.

Summary of robustness checks

Robustness check	Benchmark	Robustness result	Interpretation
Trend-augmented VAR	Joint Granger p = 0.0345	With trend: joint p = 0.6883; LHA = 0.9278, LPREC = 0.2841, LTEMP = 0.5202	Benchmark joint predictive result is trend-sensitive
Alternative Cholesky ordering	H10 FEVD: 78.47, 12.85, 1.97, 6.71%	Alternative: LPROD 78.47%, LHA 14.48%, LPREC 0.17%, LTEMP 6.87%	Shares depend on recursive identification; none is interpreted causally
Temperature in levels	Joint p with LTEMP = 0.0345	With TEMP: joint p = 0.0341; LHA = 0.2424, LPREC = 0.7365, TEMP = 0.0832	Log transformation does not drive the benchmark result; trend sensitivity remains stronger
Quandt-Andrews tests	Benchmark assumes parameter stability	Stability rejected only for LHA (p = 0.0192); not rejected for LPROD, LPREC or LTEMP	Instability is equation-specific; no common break is imposed
Bootstrap IRFs	Conventional asymptotics are weak in N = 35	2,000-repetition 95% percentile-bootstrap bands; LTEMP horizon-2 lower bound effectively zero	Climate IRF claims are weakened by bootstrap uncertainty
Without D1 and D2	Joint p = 0.0345; H10 LPREC FEVD = 1.97%	Joint p ≈ 0.0506; LTEMP p ≈ 0.0357; H10 FEVD: LPROD 75.14%, LHA 10.06%, LPREC 7.40%, LTEMP 7.41%	D1/D2 absorb part of LPREC information; precipitation results are specification-sensitive

[i] H10 = horizon 10. FEVD figures are point estimates. All robustness results are interpreted as model-conditional evidence.

[ii] Source: own elaboration.

References

Authors

Metrics

Articles in this issue

DOI: https://doi.org/10.17306/j.jard.2026.3.00022r1 | Journal eISSN: 1899-5772 | Journal ISSN: 1899-5241

Journal RSS Feed

Language: English

Page range: 348 - 360

Accepted on: Sep 7, 2026

Published on: Sep 30, 2026

Published by: Poznań University of Life Science

In partnership with: Paradigm Publishing Services

Publication frequency: 4 issues per year

Keywords:

cereal production,

harvested area,

climate variability,

Algeria,

vector autoregression,

parameter stability

Related subjects:

Business and economics,

Political economics,

Political economics, other,

Business management,

Business management, other,

Life sciences,

Life sciences, other,

Social sciences,

Political science,

Political science, other

© 2026 Siham Boudab, Nesrine Arous, Sami Bendjeddou, Hamza Ramli, published by Poznań University of Life Science
This work is licensed under the Creative Commons Attribution 4.0 License.

Volume 81 (2026): Issue 3 (September 2026)

Harvested Area, Climate Variability, and Cereal Production Dynamics in Algeria: Var Evidence, 1990–2024

Figures & Tables

Table 1.

Table 2.

Fig. 1.

Table 3.

Fig. 2.

Table 4.

Table 5.

Table 6.

Table 7.

Fig. 3.

Table 8.

Table 9.

Paradigm

My account