risk-assessment-frameworks

name	risk-assessment-frameworks
description	Political risk indicators, institutional risk, corruption risk, democratic backsliding, early warning systems for Swedish political intelligence
license	Apache-2.0

Risk Assessment Frameworks Skill

🔴 AI FIRST Quality Principle

Apply the AI FIRST principle: never accept first-pass quality. Minimum 2 iterations. Read all output, improve every section. No shortcuts.

Purpose

This skill provides comprehensive risk assessment methodologies for evaluating political, institutional, and democratic risks within the Swedish political system. It integrates international frameworks (V-Dem, Transparency International, Freedom House) with Riksdagsmonitor platform's proprietary 50+ Drools risk rules to create systematic early warning capabilities for democratic backsliding, corruption, institutional erosion, political violence, and coalition instability.

When to Use This Skill

Apply this skill when:

• ✅ Conducting democratic health assessments of Swedish institutions
• ✅ Identifying early warning signs of institutional erosion
• ✅ Assessing corruption risk at politician or party level
• ✅ Evaluating coalition stability and government sustainability
• ✅ Detecting democratic backsliding indicators
• ✅ Measuring institutional accountability effectiveness
• ✅ Analyzing political violence risk factors
• ✅ Creating risk-based intelligence priorities
• ✅ Benchmarking Sweden against international democracy standards
• ✅ Generating risk reports for stakeholders and media

Do NOT use for:

• ❌ Political persecution or targeting of legitimate opposition
• ❌ Fabricating risks to manipulate public opinion
• ❌ Undermining democratic institutions through false alarms
• ❌ Violating privacy or conducting surveillance without legal basis

Risk Assessment Framework Architecture

Integrated Risk Intelligence System

The Riksdagsmonitor platform integrates four layers of risk intelligence to create comprehensive political risk profiles:

graph TB
    subgraph "Layer 1: Data Collection"
        A1[🗳️ Behavioral Data<br/>3.5M+ votes, attendance<br/>Productivity metrics]
        A2[💰 Financial Data<br/>World Bank, ESV<br/>Economic indicators]
        A3[📊 Democracy Indices<br/>V-Dem, Freedom House<br/>International benchmarks]
        A4[📰 Media Coverage<br/>Sentiment analysis<br/>Scandal tracking]
    end
    
    subgraph "Layer 2: Risk Rules Engine (Drools)"
        A1 --> B1[Behavioral Risk Rules<br/>24 politician rules<br/>12 party rules]
        A2 --> B2[Financial Risk Rules<br/>8 corruption indicators]
        A3 --> B3[Democratic Health Rules<br/>6 institutional rules]
        A4 --> B4[Reputational Risk Rules<br/>4 scandal detection rules]
    end
    
    subgraph "Layer 3: Risk Aggregation"
        B1 --> C1[Individual Risk Profiles]
        B2 --> C2[Institutional Risk Profiles]
        B3 --> C3[Systemic Risk Profiles]
        B4 --> C4[Reputational Risk Profiles]
    end
    
    subgraph "Layer 4: Risk Intelligence"
        C1 & C2 & C3 & C4 --> D[🎯 Composite Risk Score]
        D --> E[Early Warning Alerts]
        D --> F[Risk Mitigation Strategies]
        D --> G[Intelligence Priorities]
    end
    
    style A1 fill:#e1f5ff
    style A2 fill:#e1f5ff
    style A3 fill:#e1f5ff
    style A4 fill:#e1f5ff
    style B1 fill:#fff9cc
    style B2 fill:#fff9cc
    style B3 fill:#fff9cc
    style B4 fill:#fff9cc
    style D fill:#ffe6cc
    style E fill:#ffcccc
    style F fill:#ccffcc
    style G fill:#e6ccff

1. Democratic Backsliding Detection

V-Dem Integration Framework

The Varieties of Democracy (V-Dem) project provides the world's most comprehensive democracy measurement. The Riksdagsmonitor platform integrates V-Dem indicators with behavioral data.

V-Dem Core Indicators Tracked:

• Liberal Democracy Index - Rule of law, checks on government
• Electoral Democracy Index - Free and fair elections
• Participatory Democracy Index - Citizen participation
• Deliberative Democracy Index - Quality of public discourse
• Egalitarian Democracy Index - Equal access to power

from typing import Dict, List, Tuple
import pandas as pd
import numpy as np
from datetime import datetime, timedelta

class DemocraticBackslidingDetector:
    """
    Detects democratic backsliding through trend analysis and threshold monitoring.
    
    Based on V-Dem Early Warning of Democratic Decline (Edda) methodology
    and combines international indices with Riksdagsmonitor platform behavioral data.
    """
    
    # V-Dem backsliding thresholds (0-1 scale)
    CRITICAL_THRESHOLDS = {
        'liberal_democracy_index': 0.50,  # Below = autocratization
        'electoral_democracy_index': 0.60,  # Below = electoral manipulation
        'participatory_democracy_index': 0.45,  # Below = citizen disengagement
        'deliberative_democracy_index': 0.50,  # Below = discourse degradation
        'egalitarian_democracy_index': 0.55   # Below = inequality deepening
    }
    
    def assess_democratic_health(self, country_code: str = 'SWE') -> Dict:
        """
        Comprehensive democratic health assessment for Sweden.
        
        Combines:
        1. V-Dem historical trends (5-year analysis)
        2. CIA behavioral indicators (parliamentary effectiveness)
        3. International comparison (Nordic benchmarking)
        4. Early warning signals (acceleration detection)
        """
        
        # Fetch V-Dem data
        vdem_query = 
        
        vdem_df = pd.read_sql(vdem_query, .connection, params=[country_code])
        
        
        trends = {}
         column  vdem_df.columns:
             column   [, , ]:
                X = vdem_df[].values.reshape(-, )
                y = vdem_df[column].values
                
                
                slope = np.polyfit(X.flatten(), y, )[]
                trends[column] = (slope, )
        
        
        behavioral_query = 
        
        behavioral_data = pd.read_sql(behavioral_query, .connection).iloc[]
        
        
        current_vdem = vdem_df.iloc[]
        
        
        risks = ._identify_risks(current_vdem, trends, behavioral_data)
        
        
        health_score = ._calculate_health_score(current_vdem, behavioral_data)
        
        
        early_warnings = ._detect_early_warnings(trends, current_vdem)
        
         {
            : country_code,
            : datetime.now().isoformat(),
            : {
                : (current_vdem[], ),
                : (current_vdem[], ),
                : (current_vdem[], ),
                : (current_vdem[], ),
                : (current_vdem[], )
            },
            : trends,
            : {
                : (behavioral_data[], ),
                : (behavioral_data[], ),
                : (behavioral_data[], ),
                : (behavioral_data[], ),
                : (behavioral_data[], )
            },
            : (health_score, ),
            : ._classify_health(health_score),
            : risks,
            : early_warnings,
            : ._get_nordic_comparison(current_vdem)
        }
    
     () -> []:
        
        risks = []
        
        
         indicator, threshold  .CRITICAL_THRESHOLDS.items():
             current.get(indicator, ) < threshold:
                risks.append(
                    
                    
                )
        
        
         indicator, slope  trends.items():
             slope < -:  
                risks.append(
                    
                )
        
        
         behavioral[] < :
            risks.append()
        
         behavioral[] < :
            risks.append()
        
         behavioral[] < :
            risks.append()
        
         risks  risks  []
    
     () -> :
        
        
        
        vdem_score = (
            vdem[] *  +
            vdem[] *  +
            vdem[] *  +
            vdem[] *  +
            vdem[] * 
        )
        
        
        behavioral_score = (
            (behavioral[] / ) *  +
            (behavioral[] / ) *  +
            (behavioral[] / ) * 
        )
        
         vdem_score *  + behavioral_score
    
     () -> :
        
         score >= :
             
         score >= :
             
         score >= :
             
         score >= :
             
        :
             
    
     () -> []:
        
        warnings = []
        
        
        declining_indicators = [k  k, v  trends.items()  v < -]
        
         (declining_indicators) >= :
            warnings.append(
                
            )
        
        
         (current.get(, ) <   
            trends.get(, ) < -):
            warnings.append(
                
            )
        
        
         (current.get(, ) <   
            trends.get(, ) < -):
            warnings.append(
                
            )
        
        
         (current.get(, ) <   
            trends.get(, ) < -):
            warnings.append(
                
            )
        
        
         trends.get(, ) > :
            warnings.append(
                
            )
        
         warnings  warnings  []
    
     () -> :
        
        
        query = 
        
        nordic_df = pd.read_sql(query, .connection)
        
        sweden_rank = nordic_df[
            nordic_df[] == 
        ].index[] +     nordic_df[].values  
        
         {
            :   sweden_rank  ,
            : nordic_df.to_dict()
        }

2. Corruption Risk Assessment

Transparency International Integration

The Riksdagsmonitor platform integrates Transparency International's Corruption Perceptions Index (CPI) methodology with behavioral indicators to assess corruption risk.

@Service
public class CorruptionRiskAnalyzer {
    
    /**
     * Multi-dimensional corruption risk assessment.
     * 
     * Risk dimensions:
     * 1. Financial irregularities (unexplained wealth, conflict of interest)
     * 2. Behavioral anomalies (voting patterns inconsistent with stated positions)
     * 3. Network corruption (connections to sanctioned entities)
     * 4. Transparency violations (disclosure failures, opacity)
     * 5. Accountability evasion (oversight avoidance, question dodging)
     */
    
    public CorruptionRiskProfile assessCorruptionRisk(String politicianId) {
        String sql = """
            WITH financial_risk AS (
                SELECT 
                    p.person_id,
                    
                    -- Financial disclosure completeness
                    fd.disclosure_completeness_score,
                    fd.wealth_change_unexplained_ratio,
                    fd.conflict_of_interest_declarations,
                    
                    -- Red flags
                    CASE WHEN fd.wealth_change_unexplained_ratio > 0.30 THEN 1 ELSE 0 END as wealth_anomaly_flag,
                    CASE WHEN fd.disclosure_completeness_score < 0.70 THEN 1 ELSE 0 END as disclosure_failure_flag,
                    CASE WHEN fd.conflict_of_interest_declarations = 0 AND fd.business_holdings > 0 
                         THEN 1 ELSE 0 END as coi_omission_flag
                    
                FROM person p
                LEFT JOIN financial_disclosure fd ON p.person_id = fd.person_id
                WHERE p.person_id = :politicianId
            ),
            behavioral_risk AS (
                SELECT 
                    p.person_id,
                    
                    -- Rhetoric-action gaps (potential deception)
                    raa.credibility_score,
                    raa.contradiction_count,
                    
                    -- Voting patterns (influence indicators)
                    vbs.rebel_votes,
                    vbs.total_votes,
                    
                    -- Policy area concentration (capture risk)
                    (SELECT COUNT(DISTINCT issue_category) 
                     FROM document WHERE person_id = p.person_id) as policy_focus_diversity,
                    
                    -- Red flags
                    CASE WHEN raa.credibility_score < 50 THEN 1 ELSE 0 END as credibility_flag,
                    CASE WHEN raa.contradiction_count > 20 THEN 1 ELSE 0 END as contradiction_flag
                    
                FROM person p
                LEFT JOIN rhetoric_action_alignment raa ON p.person_id = raa.person_id
                LEFT JOIN vote_ballot_summary vbs ON p.person_id = vbs.person_id
                WHERE p.person_id = :politicianId
            ),
            network_risk AS (
                SELECT 
                    p.person_id,
                    
                    -- Network connections to high-risk entities
                    COUNT(DISTINCT CASE WHEN ne.entity_risk_level = 'HIGH' 
                                       THEN ne.entity_id END) as high_risk_connections,
                    COUNT(DISTINCT CASE WHEN ne.entity_type = 'SANCTIONED_ENTITY' 
                                       THEN ne.entity_id END) as sanctioned_connections,
                    COUNT(DISTINCT CASE WHEN ne.entity_type = 'CONVICTED_CRIMINAL' 
                                       THEN ne.entity_id END) as criminal_connections,
                    
                    -- Red flags
                    CASE WHEN COUNT(DISTINCT CASE WHEN ne.entity_risk_level = 'HIGH' 
                                                  THEN ne.entity_id END) > 0 
                         THEN 1 ELSE 0 END as network_risk_flag
                    
                FROM person p
                LEFT JOIN network_entity ne ON p.person_id = ne.person_id
                WHERE p.person_id = :politicianId
                GROUP BY p.person_id
            ),
            transparency_risk AS (
                SELECT 
                    p.person_id,
                    
                    -- Response to oversight
                    oa.response_rate,
                    oa.substantive_response_rate,
                    oa.avg_response_time,
                    
                    -- Media transparency
                    COUNT(DISTINCT mi.interview_id) as media_engagement_count,
                    
                    -- Red flags
                    CASE WHEN oa.response_rate < 70 THEN 1 ELSE 0 END as evasion_flag,
                    CASE WHEN oa.substantive_response_rate < 50 THEN 1 ELSE 0 END as opacity_flag
                    
                FROM person p
                LEFT JOIN oversight_activity oa ON p.person_id = oa.person_id
                LEFT JOIN media_interview mi ON p.person_id = mi.person_id
                WHERE p.person_id = :politicianId
                GROUP BY p.person_id, oa.response_rate, oa.substantive_response_rate, 
                         oa.avg_response_time
            )
            SELECT 
                p.person_id,
                p.first_name || ' ' || p.last_name as name,
                p.party,
                
                -- Financial risk indicators
                fr.wealth_anomaly_flag,
                fr.disclosure_failure_flag,
                fr.coi_omission_flag,
                fr.wealth_change_unexplained_ratio,
                
                -- Behavioral risk indicators
                br.credibility_flag,
                br.contradiction_flag,
                br.credibility_score,
                
                -- Network risk indicators
                nr.network_risk_flag,
                nr.high_risk_connections,
                nr.sanctioned_connections,
                
                -- Transparency risk indicators
                tr.evasion_flag,
                tr.opacity_flag,
                tr.response_rate,
                
                -- Total red flags
                (fr.wealth_anomaly_flag + fr.disclosure_failure_flag + fr.coi_omission_flag +
                 br.credibility_flag + br.contradiction_flag +
                 nr.network_risk_flag +
                 tr.evasion_flag + tr.opacity_flag) as total_red_flags,
                
                -- Corruption risk score (0-100, higher = higher risk)
                (
                    (fr.wealth_anomaly_flag + fr.disclosure_failure_flag + fr.coi_omission_flag) * 8 +
                    (br.credibility_flag + br.contradiction_flag) * 6 +
                    nr.network_risk_flag * 10 +
                    (tr.evasion_flag + tr.opacity_flag) * 6 +
                    (fr.wealth_change_unexplained_ratio * 20) +
                    ((100 - br.credibility_score) / 100 * 15) +
                    (nr.high_risk_connections * 3) +
                    ((100 - tr.response_rate) / 100 * 10)
                ) as corruption_risk_score,
                
                -- Risk classification
                CASE 
                    WHEN (
                        (fr.wealth_anomaly_flag + fr.disclosure_failure_flag + fr.coi_omission_flag) * 8 +
                        (br.credibility_flag + br.contradiction_flag) * 6 +
                        nr.network_risk_flag * 10 +
                        (tr.evasion_flag + tr.opacity_flag) * 6 +
                        (fr.wealth_change_unexplained_ratio * 20) +
                        ((100 - br.credibility_score) / 100 * 15) +
                        (nr.high_risk_connections * 3) +
                        ((100 - tr.response_rate) / 100 * 10)
                    ) >= 70 THEN 'CRITICAL_CORRUPTION_RISK'
                    WHEN (
                        (fr.wealth_anomaly_flag + fr.disclosure_failure_flag + fr.coi_omission_flag) * 8 +
                        (br.credibility_flag + br.contradiction_flag) * 6 +
                        nr.network_risk_flag * 10 +
                        (tr.evasion_flag + tr.opacity_flag) * 6 +
                        (fr.wealth_change_unexplained_ratio * 20) +
                        ((100 - br.credibility_score) / 100 * 15) +
                        (nr.high_risk_connections * 3) +
                        ((100 - tr.response_rate) / 100 * 10)
                    ) >= 50 THEN 'HIGH_CORRUPTION_RISK'
                    WHEN (
                        (fr.wealth_anomaly_flag + fr.disclosure_failure_flag + fr.coi_omission_flag) * 8 +
                        (br.credibility_flag + br.contradiction_flag) * 6 +
                        nr.network_risk_flag * 10 +
                        (tr.evasion_flag + tr.opacity_flag) * 6 +
                        (fr.wealth_change_unexplained_ratio * 20) +
                        ((100 - br.credibility_score) / 100 * 15) +
                        (nr.high_risk_connections * 3) +
                        ((100 - tr.response_rate) / 100 * 10)
                    ) >= 30 THEN 'MODERATE_CORRUPTION_RISK'
                    ELSE 'LOW_CORRUPTION_RISK'
                END as risk_classification
                
            FROM person p
            LEFT JOIN financial_risk fr ON p.person_id = fr.person_id
            LEFT JOIN behavioral_risk br ON p.person_id = br.person_id
            LEFT JOIN network_risk nr ON p.person_id = nr.person_id
            LEFT JOIN transparency_risk tr ON p.person_id = tr.person_id
            WHERE p.person_id = :politicianId
            """;
        
         jdbcTemplate.queryForObject(sql, CorruptionRiskProfile.class,
            Map.of(, politicianId));
    }
}

3. Institutional Erosion Metrics

Measuring Parliamentary Effectiveness Decline

Institutional health requires effective parliamentary procedures, accountability mechanisms, and checks on executive power.

-- Institutional Erosion Index
WITH institutional_metrics AS (
    SELECT 
        -- Executive-Legislative Balance
        (SELECT AVG(oversight_effectiveness_score) 
         FROM oversight_activity 
         WHERE created_date >= NOW() - INTERVAL '2 years'
        ) as oversight_effectiveness,
        
        -- Legislative Productivity
        (SELECT COUNT(*) 
         FROM document 
         WHERE document_type = 'adopted_law' 
           AND created_date >= NOW() - INTERVAL '2 years'
        )::float / 
        (SELECT COUNT(*) 
         FROM document 
         WHERE document_type = 'adopted_law' 
           AND created_date >= NOW() - INTERVAL '4 years' 
           AND created_date < NOW() - INTERVAL '2 years'
        ) as legislative_productivity_trend,
        
        -- Committee Functionality
        (SELECT (overall_effectiveness_score) 
          committee_effectiveness 
          analysis_date  NOW()   
        )  avg_committee_effectiveness,
        
        
        ( (  avg_absent_percentage) 
          vote_ballot_summary 
          analysis_date  NOW()   
        )  avg_participation_rate,
        
        
        ( (oversight_effectiveness_score) 
          oversight_activity oa
          person p  oa.person_id  p.person_id
          p.party   ( party  government_coalition)
            oa.created_date  NOW()   
        )  opposition_effectiveness,
        
        
        ( (debate_time_allocated::  debate_time_requested) 
          parliamentary_debate 
          debate_date  NOW()   
        )  debate_time_fairness,
        
        
        ( (collaboration_intensity) 
          cross_party_collaboration 
          analysis_date  NOW()   
        )  cross_party_collaboration
),
historical_comparison  (
    
     
          metric,
        im.oversight_effectiveness  current_value,
        ( (oversight_effectiveness_score) 
          oversight_activity 
          created_date  NOW()   
            created_date  NOW()   
        )  historical_baseline,
        im.oversight_effectiveness  
        ( (oversight_effectiveness_score) 
          oversight_activity 
          created_date  NOW()   
            created_date  NOW()   
        )  change_from_baseline
     institutional_metrics im
    
     
    
     
          metric,
        im.avg_committee_effectiveness  current_value,
        ( (overall_effectiveness_score) 
          committee_effectiveness 
          analysis_date  NOW()   
            analysis_date  NOW()   
        )  historical_baseline,
        im.avg_committee_effectiveness  
        ( (overall_effectiveness_score) 
          committee_effectiveness 
          analysis_date  NOW()   
            analysis_date  NOW()   
        )  change_from_baseline
     institutional_metrics im
    
     
    
     
          metric,
        im.avg_participation_rate  current_value,
        ( (  avg_absent_percentage) 
          vote_ballot_summary 
          analysis_date  NOW()   
            analysis_date  NOW()   
        )  historical_baseline,
        im.avg_participation_rate  
        ( (  avg_absent_percentage) 
          vote_ballot_summary 
          analysis_date  NOW()   
            analysis_date  NOW()   
        )  change_from_baseline
     institutional_metrics im
)
 
    im.,
    
    
    (
          im.oversight_effectiveness    (  im.oversight_effectiveness)        
          im.legislative_productivity_trend    (  im.legislative_productivity_trend)      
          im.avg_committee_effectiveness    (  im.avg_committee_effectiveness)        
          im.avg_participation_rate    (  im.avg_participation_rate)        
          im.opposition_effectiveness    (  im.opposition_effectiveness)        
          im.debate_time_fairness    (  im.debate_time_fairness)        
          im.cross_party_collaboration    (  im.cross_party_collaboration)       
    )  institutional_erosion_index,
    
    
     
         (
              im.oversight_effectiveness    (  im.oversight_effectiveness)        
              im.legislative_productivity_trend    (  im.legislative_productivity_trend)      
              im.avg_committee_effectiveness    (  im.avg_committee_effectiveness)        
              im.avg_participation_rate    (  im.avg_participation_rate)        
              im.opposition_effectiveness    (  im.opposition_effectiveness)        
              im.debate_time_fairness    (  im.debate_time_fairness)        
              im.cross_party_collaboration    (  im.cross_party_collaboration)       
        )    
         (
              im.oversight_effectiveness    (  im.oversight_effectiveness)        
              im.legislative_productivity_trend    (  im.legislative_productivity_trend)      
              im.avg_committee_effectiveness    (  im.avg_committee_effectiveness)        
              im.avg_participation_rate    (  im.avg_participation_rate)        
              im.opposition_effectiveness    (  im.opposition_effectiveness)        
              im.debate_time_fairness    (  im.debate_time_fairness)        
              im.cross_party_collaboration    (  im.cross_party_collaboration)       
        )    
         (
              im.oversight_effectiveness    (  im.oversight_effectiveness)        
              im.legislative_productivity_trend    (  im.legislative_productivity_trend)      
              im.avg_committee_effectiveness    (  im.avg_committee_effectiveness)        
              im.avg_participation_rate    (  im.avg_participation_rate)        
              im.opposition_effectiveness    (  im.opposition_effectiveness)        
              im.debate_time_fairness    (  im.debate_time_fairness)        
              im.cross_party_collaboration    (  im.cross_party_collaboration)       
        )    
         
      erosion_classification,
    
    
    ( 
         
             (  change_from_baseline     )   
                 
             (  change_from_baseline     )   
                 
             (  change_from_baseline     )   
                 
             
        
      historical_comparison
    )  historical_trend
    
 institutional_metrics im;

4. Coalition Instability Prediction

Government Sustainability Forecasting

Coalition governments in parliamentary systems are vulnerable to collapse. The Riksdagsmonitor platform predicts coalition stability.

from sklearn.ensemble import GradientBoostingClassifier
from typing import Dict, List
import pandas as pd

class CoalitionStabilityPredictor:
    """
    Predicts coalition stability and government sustainability.
    
    Features:
    - Intra-party discipline (deviation rates)
    - Inter-party alignment (voting agreement)
    - Policy conflict indicators (deviation on key issues)
    - Leadership approval ratings
    - Economic conditions
    - Scandal/crisis events
    - Time in office (fatigue factor)
    """
    
    def __init__(self):
        self.model = GradientBoostingClassifier(n_estimators=200, max_depth=5)
        self.trained = False
    
    def predict_stability(
        self, 
        coalition_parties: List[str],
        prediction_horizon_months: int = 12
    ) -> Dict:
        """
        Predicts coalition stability over specified time horizon.
        
        Returns:
        - Survival probability (0-1)
        - Key risk factors
        - Collapse scenarios
        - Recommended monitoring priorities
        """
        
        # Extract coalition features
        query = """
        WITH coalition_features AS (
            SELECT 
                -- Party discipline
                AVG(pd.avg_deviation_rate) as avg_intra_party_deviation,
                MAX(pd.max_deviation_rate) as max_intra_party_deviation,
                STDDEV(pd.avg_deviation_rate) as deviation_heterogeneity,
                
                -- Cross-party alignment
                AVG(cpa.alignment_rate) as avg_cross_party_alignment,
                MIN(cpa.alignment_rate) as min_cross_party_alignment,
                
                -- Policy conflict indicators
                COUNT(DISTINCT CASE 
                    WHEN pd.issue_category IN ('economic_policy', 'foreign_policy', 'justice')
                        AND pd.avg_deviation_rate > 15 
                    THEN pd.issue_category 
                END) as critical_policy_conflicts,
                
                -- Leadership factors
                AVG(lp.approval_rating) as avg_leadership_approval,
                MIN(lp.approval_rating) as min_leadership_approval,
                
                -- Time factors
                EXTRACT(MONTH FROM NOW() - MIN(gc.formation_date)) as months_in_office,
                
                -- Crisis events
                COUNT(DISTINCT ce.crisis_id) as recent_crises,
                
                -- Scandal exposure
                COUNT(DISTINCT se.scandal_id) as recent_scandals
                
            FROM party_deviation pd
            JOIN cross_party_alignment cpa ON pd.party IN (cpa.party_a, cpa.party_b)
            JOIN leadership_profile lp ON pd.party = lp.party
            JOIN government_coalition gc ON pd.party = gc.party
            LEFT JOIN crisis_event ce ON ce.event_date >= NOW() - INTERVAL '6 months'
            LEFT JOIN scandal_event se ON se.event_date >= NOW() - INTERVAL '6 months'
                                        AND se.party IN (SELECT unnest(%s))
            WHERE pd.party = ANY(%s)
                AND pd.analysis_date >= NOW() - INTERVAL '6 months'
            GROUP BY 1, 2, 3, 4, 5, 6, 7, 8, 9, 10, 11
        )
        SELECT * FROM coalition_features
        """
        
        features = pd.read_sql(
            query, 
            .connection, 
            params=[coalition_parties, coalition_parties]
        ).iloc[]
        
          .trained:
            .train()  
        
        
        X = ._prepare_features(features)
        
        
        survival_probability = .model.predict_proba(X)[][]
        
        
        risk_factors = ._identify_risk_factors(features)
        
        
        scenarios = ._generate_scenarios(features, survival_probability)
        
         {
            : coalition_parties,
            : prediction_horizon_months,
            : (survival_probability, ),
            : ._classify_stability(survival_probability),
            :   (survival_probability - ) >   ,
            : risk_factors,
            : scenarios,
            : ._recommend_monitoring(features)
        }
    
     () -> []:
        
        risks = []
        
         features[] > :
            risks.append({
                : ,
                : ,
                : (features[], ),
                : 
            })
        
         features[] < :
            risks.append({
                : ,
                : ,
                : (features[], ),
                : 
            })
        
         features[] > :
            risks.append({
                : ,
                : ,
                : (features[]),
                : 
            })
        
         features[] < :
            risks.append({
                : ,
                : ,
                : (features[], ),
                : 
            })
        
         features[] > :
            risks.append({
                : ,
                : ,
                : (features[]),
                : 
            })
        
         features[] > :
            risks.append({
                : ,
                : ,
                : (features[]),
                : 
            })
        
         (risks, key= x: 
                     {: , : , : }.get(x[], ),
                     reverse=)
    
     () -> []:
        
        scenarios = []
        
        
         features[] > :
            scenarios.append({
                : ,
                : ,
                : (
                    base_probability * ( + features[] * ),
                    
                ),
                : ,
                : [
                    ,
                    ,
                    
                ]
            })
        
        
         features[] < :
            scenarios.append({
                : ,
                : ,
                : (
                    base_probability * ( + ( - features[]) / ),
                    
                ),
                : ,
                : [
                    ,
                    ,
                    
                ]
            })
        
        
         features[] > :
            scenarios.append({
                : ,
                : ,
                : (
                    base_probability * ( + features[] / ),
                    
                ),
                : ,
                : [
                    ,
                    ,
                    
                ]
            })
        
        
         features[] > :
            scenarios.append({
                : ,
                : ,
                : (
                    base_probability * ( + features[] * ),
                    
                ),
                : ,
                : [
                    ,
                    ,
                    
                ]
            })
        
         (scenarios, key= x: x[], reverse=)
    
     () -> :
        
         probability >= :
             
         probability >= :
             
         probability >= :
             
        :
             
    
     () -> []:
        
        priorities = []
        
         features[] < :
            priorities.append()
        
         features[] < :
            priorities.append()
        
         features[] > :
            priorities.append()
        
         features[] > :
            priorities.append()
        
         features[] > :
            priorities.append()
        
         priorities  priorities  [
            
        ]

5. Political Violence Risk Indicators

Early Warning System for Political Violence

Political violence threatens democratic stability. The Riksdagsmonitor platform monitors behavioral and contextual indicators.

-- Political Violence Risk Assessment
WITH violence_risk_indicators AS (
    SELECT 
        -- Rhetorical escalation
        COUNT(CASE WHEN dc.contains_violent_rhetoric = TRUE THEN 1 END) as violent_rhetoric_count,
        COUNT(CASE WHEN dc.contains_dehumanizing_language = TRUE THEN 1 END) as dehumanization_count,
        COUNT(CASE WHEN dc.contains_threat = TRUE THEN 1 END) as threat_count,
        
        -- Polarization indicators
        AVG(pp.polarization_index) as avg_polarization,
        MAX(pp.polarization_index) as max_polarization,
        
        -- Protest activity
        COUNT(DISTINCT pe.protest_event_id) as protest_count,
        AVG(pe.violence_level) as avg_protest_violence,
        COUNT(CASE WHEN pe.violence_level >= 3 THEN 1 )  violent_protests,
        
        
        ( ()  hate_crime_incident 
          incident_date  NOW()   
            political_motivation  
        )  political_hate_crimes,
        
        
        ( oec.extremist_content_id)  extremist_content_items,
        
        
        (  ma.incitement_score     )  high_incitement_articles
        
     document_content dc
     party_polarization pp  
      protest_event pe  pe.event_date  NOW()   
      online_extremist_content oec  oec.detected_date  NOW()   
      media_article ma  ma.published_date  NOW()   
     dc.created_date  NOW()   
)
 
    vri.,
    
    
    (
        LEAST(vri.violent_rhetoric_count  , )   
        LEAST(vri.dehumanization_count  , )   
        LEAST(vri.threat_count  , )   
        vri.avg_polarization   
        LEAST(vri.violent_protests  , )   
        LEAST(vri.political_hate_crimes  , )   
        LEAST(vri.extremist_content_items  , )  
    )    violence_risk_score,
    
    
     
         (
            LEAST(vri.violent_rhetoric_count  , )   
            LEAST(vri.dehumanization_count  , )   
            LEAST(vri.threat_count  , )   
            vri.avg_polarization   
            LEAST(vri.violent_protests  , )   
            LEAST(vri.political_hate_crimes  , )   
            LEAST(vri.extremist_content_items  , )  
        )      
         (
            LEAST(vri.violent_rhetoric_count  , )   
            LEAST(vri.dehumanization_count  , )   
            LEAST(vri.threat_count  , )   
            vri.avg_polarization   
            LEAST(vri.violent_protests  , )   
            LEAST(vri.political_hate_crimes  , )   
            LEAST(vri.extremist_content_items  , )  
        )      
         (
            LEAST(vri.violent_rhetoric_count  , )   
            LEAST(vri.dehumanization_count  , )   
            LEAST(vri.threat_count  , )   
            vri.avg_polarization   
            LEAST(vri.violent_protests  , )   
            LEAST(vri.political_hate_crimes  , )   
            LEAST(vri.extremist_content_items  , )  
        )      
         
      risk_classification,
    
    
     
         vri.threat_count    vri.violent_protests   
             
         
      immediate_action_required
    
 violence_risk_indicators vri;

ISMS Compliance Mapping

ISO 27001:2022 Controls

Control	Risk Assessment Application
A.5.7 - Threat intelligence	Systematic threat intelligence from risk frameworks
A.5.10 - Acceptable use of information and other associated assets	Ethical use of political risk intelligence
A.8.16 - Monitoring activities	Continuous risk monitoring and early warning

NIST Cybersecurity Framework 2.0

Function	Risk Assessment Integration
IDENTIFY (ID.RA)	Comprehensive political risk identification
DETECT (DE.CM)	Early warning detection systems
RESPOND (RS.AN)	Risk-based response prioritization

CIS Controls v8

Control	Application
CIS Control 4 - Secure Configuration	Secure risk assessment system configuration
CIS Control 8 - Audit Log Management	Risk assessment audit trail

Hack23 ISMS Policy References

This skill implements requirements from:

• Secure Development Policy - Risk assessment methodology standards
• Information Security Policy - Ethical risk assessment practices
• Risk Assessment Methodology - Risk calculation methods
• Threat Modeling - Threat-based risk assessment

References

Risk Assessment Literature

1. Coppedge, M., et al. (2021). V-Dem Codebook v11. Varieties of Democracy (V-Dem) Project.
2. Transparency International (2023). Corruption Perceptions Index: Methodology.
3. Lührmann, A., & Lindberg, S. I. (2019). "A Third Wave of Autocratization is Here: What is New About It?" Democratization, 26(7), 1095-1113.
4. Schedler, A. (2013). The Politics of Uncertainty: Sustaining and Subverting Electoral Authoritarianism. Oxford University Press.

Database Intelligence Sources

• DATABASE_VIEW_INTELLIGENCE_CATALOG.md - Complete view documentation
• RISK_RULES_INTOP_OSINT.md - 50+ risk rule specifications
• DATA_ANALYSIS_INTOP_OSINT.md - Analysis frameworks
• INTELLIGENCE_DATA_FLOW.md - Risk data flow mapping

---

🔗 Integration with agentic workflows & analysis artifacts

This skill is consumed by the 11 agentic news workflows in .github/workflows/news-*.md. The authoritative contract lives in .github/prompts/README.md; this skill supplies domain expertise on top of that contract.

• Analysis product → ai-driven-analysis-guide.md + every template in analysis/templates/.
• Required before any article: 9 core artifacts (14 for Tier-C) in analysis/daily/$ARTICLE_DATE/$SUBFOLDER/; 05-analysis-gate.md is the single blocking gate.
• gh-aw v0.69.3 docs: abridged · complete · blog series.

---

🌐 IMF as Primary Source for Economic Risk Indicators

Effective: 2026-04-24

IMF risk-indicator matrix

Risk dimension	IMF source	Indicator(s)	Threshold of concern
Sovereign debt sustainability	IMF WEO + Fiscal Monitor	GGXWDG_NGDP (general government debt %GDP), GGXWDN_NGDP (net debt)	>60% (Maastricht) / >90% (Rogoff–Reinhart)
Fiscal deficit	IMF WEO + FM	GGXCNL_NGDP (overall balance), GGSB_NPGDP (cyclically-adjusted)	<-3% (Maastricht), <-1% structural
External vulnerability	IMF WEO + BOP	BCA_NGDPD (current account %GDP), reserves coverage	<-4% (warning), <-6% (acute)
Inflation pressure	IMF WEO + IFS	PCPIPCH (CPI inflation)	>Riksbank target +200bps
Growth slowdown	IMF WEO	NGDP_RPCH (real GDP growth, T+5 projections)	<0% any year in 5-year horizon
Trade exposure	IMF DOTS	Bilateral concentration index	top-3 partner share >50%
Commodity-price exposure	IMF PCPS	Energy/food index volatility	rolling stdev >threshold

Canonical risk rule: Every economic-risk assessment in Riksdagsmonitor uses IMF as the primary source for the risk indicator and IMF projections (T+5) for the forward-looking risk horizon. World Bank governance indicators (WGI) supplement for political-risk dimension. See .github/aw/ECONOMIC_DATA_CONTRACT.md v2.1 and analysis/imf/agentic-integration.md.

---

🔭 Horizon stratification

Authoritative source: .github/prompts/ext/long-horizon-forecasting.md. Runtime helper: scripts/horizon-context.ts.

Risk assessments use horizon bands to calibrate threat-likelihood language:

Band	Days	WEP language ceiling	Risk-assessment application
72h	3	very likely / very unlikely	Imminent threat (e.g. no-confidence motion filed)
week	7	likely / unlikely	Near-term policy risk, scheduled government decisions
month	30	likely / unlikely	Budget-cycle risk, regulatory deadline exposure
quarter	90	roughly even / about even	Session-level institutional stability
year	365	roughly even; stronger requires ≥ 3 cycle-aged sources	Fiscal sustainability (IMF WEO/FM T+1–T+5), structural reform risk
cycle	1460	roughly even / unlikely; never likely without ≥ 3 cycle-aged sources	Long-range political-risk trajectory, democratic-backsliding indicators
election	1460	scenario-driven; coalition outcomes never above "likely"	Election-cycle regime-change risk, coalition-formation scenarios

Cycle-rollover rules (±30 days of election anchor) are defined in .github/prompts/ext/cycle-rollover.md.