Table 3
Flows that Contain Endogenous and Exogenous Components.
|
Description |
Variable |
|
Flow from employment to unemployment; job destruction |
FEUV |
|
Total flow from unemployment to employment |
FUE |
|
Total outflow from unemployment |
UO |
|
Total outflow from employment |
EO |
|
Total inflow to unemployment |
UI |
|
Total inflow to employment |
EI |
Table 4
Numerical values for baseline model (*=Number of persons/jobs x 1000, annual
averages).
|
Variable |
Value |
Variable |
Value |
|
U* |
750 |
Feu * |
600 |
|
FENV * |
300 |
Fuev * |
200 |
|
FNEV * |
400 |
Feuv * |
300 |
|
E* |
8,600 |
UO * |
4,600 |
|
V* |
75 |
r |
0.03 |
|
θ |
0.9 |
c |
0.9 |
|
α |
0.5 |
EO * |
4,200 |
|
V I* |
600 |
τ |
0.08 |
|
VOU* |
200 |
У |
1 |
|
VOEX* |
400 |
b |
0.3wss |
rate to 5% and has no participants in ALMPs. This unemployment target is
a substantial reduction from the baseline rate of nearly 8%, and is a target
that will be difficult for any method to achieve.
The input to the model (or control) is the proportion of the long-term un-
employed on ALMPs (λ). The control target is zero. There is no constraint
within the model on this proportion so the control method should incorporate
a realistic constraint. We use the constraint of 0 ≤ λ ≤ 0.5 so that a maxi-
mum of 50% of the long-term unemployed can participate in ALMPs in any
one month.
We follow the convention of using a quadratic objective function where the
optimization aims to reduce the sum of squared deviations from the output
(unemployment) and control (λ) targets. The deviations are weighted and we