Compute Temporal Employment Indicators from vecshift Data

Computes comprehensive employment indicators (employment rate, contract quality, wage statistics, stability measures, and transition patterns) at quarterly or monthly intervals from vecshift-processed employment data. Supports flexible aggregation and consolidation options for robust impact evaluation analysis.

Usage

compute_temporal_employment_indicators(
  data,
  period_type = "quarterly",
  indicators = c("employment_rate", "contract_quality", "stability_index"),
  id_var = "cf",
  start_date_var = "inizio",
  end_date_var = "fine",
  duration_var = "durata",
  contract_type_var = "COD_TIPOLOGIA_CONTRATTUALE",
  employment_intensity_var = "prior",
  employment_var = "arco",
  wage_var = NULL,
  over_id_var = "over_id",
  use_consolidation = TRUE,
  reference_date = NULL,
  employment_threshold = 1,
  quality_weights = NULL,
  group_vars = NULL,
  output_format = "long",
  verbose = FALSE
)

Arguments

data

data.table containing vecshift-processed employment records

period_type

Character. Temporal aggregation: "quarterly" or "monthly". Default: "quarterly"

indicators

Character vector. Indicators to compute. Options:

"employment_rate": Proportion of time employed in each period
"contract_quality": Duration-weighted average contract quality (0-1 scale)
"wage_statistics": Average wages, wage growth (when available)
"stability_index": Employment stability and contract duration metrics
"contract_types": Distribution of contract types by period
"transition_patterns": Employment state transitions between periods

Default: c("employment_rate", "contract_quality", "stability_index")

id_var

Character. Individual identifier column name. Default: "cf"

start_date_var

Character. Contract start date column. Default: "inizio"

end_date_var

Character. Contract end date column. Default: "fine"

duration_var

Character. Contract duration column (days). Default: "durata"

contract_type_var

Character. Contract type code column. Default: "COD_TIPOLOGIA_CONTRATTUALE"

employment_intensity_var

Character. Employment intensity column (1=full-time). Default: "prior"

employment_var

Character. Employment status column (arco >= 1 = employed). Default: "arco"

wage_var

Character. Wage/salary column. Default: NULL (auto-detect)

over_id_var

Character. Consolidation identifier from vecshift. Default: "over_id"

use_consolidation

Logical. Use over_id for period consolidation? Default: TRUE

reference_date

Date. Reference date for period numbering. Default: NULL (min date)

employment_threshold

Numeric. Minimum days to count as employed in period. Default: 1

quality_weights

DEPRECATED. This parameter is ignored. Contract quality is now calculated using Kaplan-Meier survival analysis to determine median survival duration for each contract type, providing data-driven quality scores (0.3-1.0) rather than arbitrary fixed weights. This scientific approach ensures quality metrics reflect actual employment durability patterns in your data.

group_vars

Character vector. Variables for subgroup analysis. Default: NULL

output_format

Character. "long" (period-person rows) or "wide" (period columns). Default: "long"

verbose

Logical. Print progress information? Default: FALSE

Value

data.table with temporal employment indicators:

cf: Individual identifier
period: Temporal period number (1, 2, 3, ...)
period_start/period_end: Period date range
employment_rate: Proportion of period employed (0-1)
contract_quality: Duration-weighted average quality (0-1)
avg_wage: Average wage when employed (if available)
stability_index: Employment stability score (0-1)
n_contracts: Number of contracts in period
total_days_employed: Total employment days in period
pct_permanent: Proportion of time in permanent contracts
transition_from/transition_to: Employment transitions (if requested)

Additional columns depend on requested indicators and group variables.

Details

Period Calculation: Periods start from the reference date (default: earliest contract date). Quarterly periods span ~90 days, monthly periods ~30 days. Contracts spanning multiple periods contribute proportionally to each period.

Consolidation: When use_consolidation=TRUE, overlapping contracts with same over_id are consolidated before analysis, providing cleaner employment episode identification.

Employment Rate: Calculated as days_employed / period_duration, capped at 1.0 for periods with overlapping contracts. Minimum employment threshold filters noise.

Contract Quality: Calculated using Kaplan-Meier survival analysis to determine median survival duration for each contract type. Contract types with longer median survival times receive higher quality scores (0.3-1.0 scale), where 1.0 represents the most durable contract types and 0.3 the least durable. This empirical, data-driven approach replaces arbitrary fixed weights (e.g., permanent contracts = 1.0, temporary = 0.5) with actual employment durability metrics derived from survival analysis of the employment data itself.

Stability Index: Combines contract duration consistency, employment continuity, and transition frequency into a 0-1 stability measure.

Examples

if (FALSE) { # \dontrun{
# Load sample vecshift data
employment_data <- readRDS("data/sample.rds")

# Basic quarterly employment indicators with survival-based contract quality
quarterly_indicators <- compute_temporal_employment_indicators(
  data = employment_data,
  period_type = "quarterly",
  indicators = c("employment_rate", "contract_quality", "stability_index")
)
# Note: contract_quality automatically uses survival analysis to determine
# quality scores based on median survival duration by contract type

# Monthly indicators with wage analysis
monthly_indicators <- compute_temporal_employment_indicators(
  data = employment_data,
  period_type = "monthly", 
  indicators = c("employment_rate", "contract_quality", "wage_statistics"),
  wage_var = "retribuzione_media"
)

# Comprehensive analysis with transitions
comprehensive_indicators <- compute_temporal_employment_indicators(
  data = employment_data,
  indicators = c("employment_rate", "contract_quality", "stability_index", 
                "contract_types", "transition_patterns"),
  group_vars = c("gender", "education"),
  output_format = "wide"
)

# Examine survival-based quality scores by contract type
# This shows how different contract types are ranked by durability
quality_by_type <- comprehensive_indicators[, .(
  avg_quality = mean(avg_contract_quality, na.rm = TRUE),
  median_quality = median(avg_contract_quality, na.rm = TRUE)
), by = contract_type][order(-avg_quality)]
print(quality_by_type)  # Higher scores = more durable contract types

# Integration with impact evaluation
treatment_assignment <- data.table(
  cf = unique(employment_data$cf),
  treatment_group = sample(c("control", "training", "subsidy"), 
                          length(unique(employment_data$cf)), replace = TRUE)
)

indicators_with_treatment <- merge(quarterly_indicators, treatment_assignment, by = "cf")

# Difference-in-differences analysis
did_analysis <- difference_in_differences(
  data = indicators_with_treatment,
  outcome_vars = c("employment_rate", "contract_quality", "stability_index"),
  treatment_var = "treatment_group",
  time_var = "period", 
  id_var = "cf"
)
} # }

Usage

Arguments

Value

Details

See also

Examples