Revealed comparative advantage of skills by profession and window

Computes, for each window, the posting-based revealed comparative advantage of skill \(k\) in profession \(j\).

Usage

compute_rca_panel(
  panel,
  skill_col = "skill_id",
  prof_col,
  time_col = "mese_idx",
  x_col = "x",
  n_col = "n",
  windows
)

Arguments

panel

Profession-level share panel (e.g. from compute_share_panel() with group_cols = prof_col).

skill_col

Skill column. Default "skill_id".

prof_col

Profession column.

time_col, x_col, n_col

See compute_share_trend().

windows

Named list of numeric vectors c(from, to), e.g. list(W1 = c(1, 12), W3 = c(25, 36)).

Value

A data.table with finestra, skill_col, prof_col, x, n, quota (\(s_{kj}\)), quota_skill (\(s_k\)) and rca.

Details

Within window \(W\), \(s_{kj} = \sum_{t \in W} x_{kjt} / \sum_{t \in W} n_{jt}\) is the share of the profession's postings that mention the skill and \(s_k = \sum_j \sum_{t} x_{kjt} / \sum_j \sum_t n_{jt}\) the share over all professions. The index is \(RCA_{kj} = s_{kj} / s_k\); values of at least 1 indicate specialisation. The denominator of a profession-period must be the same on all its rows; professions are assumed to partition the postings.

Examples

panel <- data.table::data.table(
  cp4 = rep(c("p1", "p2"), each = 2), skill_id = rep(c("a", "b"), 2),
  mese_idx = 1, x = c(30, 5, 10, 20), n = rep(c(100, 100), each = 2)
)
compute_rca_panel(panel, prof_col = "cp4", windows = list(W1 = c(1, 1)))
#>    finestra skill_id    cp4     x     n quota quota_skill   rca
#>      <char>   <char> <char> <num> <num> <num>       <num> <num>
#> 1:       W1        a     p1    30   100  0.30       0.200   1.5
#> 2:       W1        b     p1     5   100  0.05       0.125   0.4
#> 3:       W1        a     p2    10   100  0.10       0.200   0.5
#> 4:       W1        b     p2    20   100  0.20       0.125   1.6