Poverty and Employment in Timber Dependent Counties

1
Poverty and Employment in Timber Dependent
Counties

• By Peter Berck, Christopher Costello, Sandra
  Hoffmann and Louise Fortmann

2
The ESA

• "The loss is evident in the lines at the soup
  kitchens. And the loss is evident in the homes
  where unemployed workers, anxious, depressed,
  sunk in despair, lash out at their loved ones or
  find solace in alcohol or drugs.
• Archbishop Thomas Murphy

3
Decrease in Cutting Timber

• Endangered Species and Running Out
• Spotted Owl, Marbled Murrelet, Many Salmonids
• Just plain ran out of big trees to cut
• And then came the spotted owl, and almost
  overnight the hauling jobs dried up and we had
  our electricity turned off and finally we
  received a foreclosure notice on this farm

4
(No Transcript)
5
Mobility and Determination

• Does cutting really drive employment or poverty?
• Mobility of labor
• Changing jobs
• State level variables could determine

6
Testable propositions

• County poverty or timber jobs are not part of any
  long run relationship with each other.
• County poverty is weakly exogenous (not
  determined by timber jobs)
• State variables determine the level of both
  poverty and jobs

7
Does Cutting Trees

• Reduce Poverty in Timber dependent counties?
• Increase Employment
• by more than one job for each new timber job?
• by one job or less?

8
Two Modeling Philosophies

• CGE/IO/SAM multiplier models
• capture all relevant economics
• assumptions on difficult to measure parameters
  can drive results
• labor mobility (no real migration data)
• openness to trade (interstate trade unmeasured)
• relation of product to labor input (product
  unmeasured)

9
The Error Correction VAR

• Nearly no economics imposed on model
• Uses available real data
• Cant explain why
• but Can measure impact multipliers
• and be used to find Long Run Relationships

10
Form of Cointegrating Equation

• () Dyt f G1Dyt-1 Gk-1Dyt-k1 Pyt-1
  ?Dt et,
• number of coint vectors is rank of P
• is also number of long run relations
• some variables can be excluded from long run
  relations
• some variables dont adjust to LR rel.

11
The Data

• Monthly from 1984-1993
• County
• timber employment
• non timber employment
• AFDC UP caseload
• State
• timber employment
• non timber employment

12
Model for each County

• Johansens MLE of the possibly cointegrated VAR
  using the 3 county and 2 state variables.
• lag length
• rank of cointegrating space
• coefficient estimates, incl. coint vectors
• Exclusion and Weak Exogeneity tests
• Calculate SR impact multipliers

13
(No Transcript)
14
Results
15
Timber and Poverty LR

• Exclusion of Timber or Poverty
• Not Poverty Humboldt, Siskyou
• Not Timber Trinity, Del Norte, Amador,
  Tuolumne.
• Poverty Weakly exogenous
• Plumas, Mendocino
• Increase Timber, INCREASE poverty
• Tehama

16
Poverty Conclusion

• Rank 3 Stabile povery unless state level
  variables change
• Shasta
• Only in Lassen of the 11 counties may timber
  employ reduce AFDC-UP in LR

17
Timber Jobs Special?

• Job is a Job
• In four of 11 counties timber jobs shift
  cointegrating space same as any other job.
  Poverty same.
• 100 new timber jobs 78 jobs 2 years later.
  Mult is less than 1!

18
SR timber Multipliers

• 100 new timber jobs 3 less cases of AFDC-UP or
• 1 timber jobs increase 14/100 poverty
  decrease
• Non timber employment does fractionally better

19
Conclusion

• Cutting more trees wont do anything for poverty
  in the LR and very little in the SR.
• Employment Cutting more trees doesnt have base
  like multipliers.