Conservation Policy Impacts in Tropical Dry Forest: regionally

1
Conservation Policy Impactsin Tropical Dry
Forest regionally spatially focused policy
analyses(given other social/natural drivers of
land use)
An SGP-HD project, in collaboration with CRN2-021
Tropi Dry.
2
Four Partners

• Arturo Sanchez U. of Alberta P.I. of Tropi
  Dry
• (deforestation collaboration since 1997,
  integrated)
• Juan Robalino LACEEP EfD (LatAm), Co-PI
• (deforestation collaboration since 2000,
  evaluation)
• Fundacion Neotropica Carlos Leon et al.
• Universidade de Montes Claros Mario Marcos et al
• (following Tropi Dry work in their particular
  sites)

3
Four Questions

• 1. Does the clearing dynamic within tropical dry
  forests differ from dominant forces for other
  types of forest?
• 2. How effective have the existing conservation
  policies been in reducing deforestation in
  tropical dry forest?
• 3. How do local Tropi-Dry cases exemplify /or
  differ from the averages revealed in a regression
  analysis?
• 4. Results robust to improved Tropi-Dry remote
  sensing?

4
HOW HAS THE SCIENCE DEVELOPED?

• 3. Tropi-Dry cases held hostage by
  moves/paperwork
• 4. Results robustness will have to wait for
  the results
• Q1. Does the clearing dynamic within tropical dry
  forests differ from dominant forces for other
  types of forest?
• A1. YES -- BUT WORK ON POLICY IMPLICATIONS
• Q2. How effective have the existing conservation
  policies been in reducing deforestation? SEE
  PRIOR NEW

5
Empirical Issue policy location is biased

• much policy discussion less empirical evaluation
  little attention to where the policies are
  implemented
• where is determined by agency land owner
  choices which we argue likely respond to factors
  we observe
• empirically examine the implications of those
  choices
• - LOCATIONS REALLY LOOK NON-RANDOM
• - THIS REALLY MATTERS FOR EVALUATION

6
Find Apples to Apples Good Comparisons
admit when cannot do so, e.g. below for
ecopayments
7
Bias Policy Issue additionality is overstated

• much attention to observing outcomes with
  policies little evaluation of whether those
  imply any impact
• impact is determined not by forest outcomes
  alone must ask if differ from what we see
  without policy
• empirically, we improve estimates of no-policy
  case
• - POLICY LOCATIONS FACE LOWER PRESSURE
• - COMPARE POLICY APPLES TO OTHER APPLES LOWER
  IMPACT ESTIMATES BY AT LEAST 2/3

8
Additionality -- policy implication Priorities
(GRUAS figure from Pfaff Sanchez-Azofeifa 2004)
9
Park Impacts (regional/spatial) San Jose far
Park Effects on 86-97 Deforestation, N 4 in
each method NOW ONLY FAR FROM SAN JOSE i.e.
over 85km (562 observations)
Adj. Diff. in Means
Difference in Means
Strategy
-1.99
-9.38
Using All of the Untreated (Naive)
-0.60
-0.03
Propensity Score Matching (PSM)
-0.61
-0.88
Covariate Matching (CM)
The less of a threat in the absence of parks, on
average the less impact on deforestation from
even a perfectly protected park. This is one
factor to add within siting.
10
Park Impacts (regional/spatial) San Jose close
Park Effects on 86-97 Deforestation, N 4 in
each method NOW ONLY CLOSE TO SAN JOSE i.e.
under 85km (396 observations)
Adj. Diff. in Means
Difference in Means
Strategy
-1.99
-9.38
Using All of the Untreated (Naive)
-5.14
-0.72
Propensity Score Matching (PSM)
-2.75
-3.97
Covariate Matching (CM)
Naturally, the existing species and the cost of
the land probably also differ for areas closer to
a national capitol. However, along with those
data, consider parks impact.
11
HOW PLANNING TO COMMUNICATE TO POLICY,
STAKEHOLDER COMMUNITY ?

• Ecopayments Average Impacts visited FONAFIFO
  first
• Payments Parks (averages) both GEF working
  papers
• Education January 2008 LACEEP short course on
  impact
• Dissemination January 2008 Moore Foundation
  workshop on economics in conservation (note CI
  talk last week)
• Extensions in Mexico, talking with agencies
  personnel

12
INTERDISC. OPPORTUNITY/OBSTACLES ?

• Prior Deforestation Teams (Tinker, NSF, NCEAS)
• understanding the inputs, processing, outputs of
  other fields
• thinking about varied requirements for CVs and
  career paths
• opportunities output more policy relevant
  disseminate too
• Even Then, Still A Challenge To Work Across
  Scales
• policy evaluation can gain from stepping back to
  high level
• yet relationships need to have meaning for
  specific sites too
• possible because CRN2 Tropi Dry is already at a
  site level