Evaluating an experimental
catch shares program for Gulf
of Mexico headboats:
Early returns
Daniel Willard
Joshua Abbott
Arizona State University Environmental Defense Fund
Oceans Program
jkabbott.faculty.asu.edu
What are we evaluating?
• Allocation of trips and
landings and landings
over the season
• Allocation of fish per
customer
• Customers per trip
• Discards
• Revenues and profits
• Welfare gains to anglers
from greater potential
trip quality over a longer
season
Data
1. 2003-2014 logbook data for vessels owned by GHC
members (2003-2014)
2. Aggregated 2003-2014 logbook data by
region/year (TX, LA/MS, AL, NWFL, SWFL)
3. Partially disaggregated logbook data at the
region/week scale for non-GHC vessels
o
Data censored if <3 vessels present by region/week for confidentiality
4. Data on trip pricing and costs from headboat
owners
5. Onboard survey of GHC customers and follow-up
internet stated preference/revealed preference
survey
Anglers on GHC vessels
Policy evaluation:
what’s the counterfactual?
• Ideal comparison: “what happened under the EFP
relative to what would have happened under the most
policy-relevant baseline without the EFP?”
o But what is the “policy relevant baseline”?
o What is the right empirical approach to impute the missing counterfactual?
• Preliminary comparisons for today:
o GHC vessels vs. non-GHC pre-2014 vs. post-2014 (diff-in-diff)
• Currently very limited by aggregation of data for non-GHC vessels
• Concerns: are non-GHC vessels indirectly treated? What about the very
short 2014 season?
• More to come…
o GHC vessels in 2014 vs. pre-2014 (first difference)
• The bulk of this presentation
• None of what we present today should be interpreted as
the “causal effect” of the EFP (yet)
I. GHC vs. non-GHC:
comparisons across time
Number of Total Angler-Days
AL
NWFL
TX
ALL
SWFL
150000
100000
50000
13
20
11
20
09
20
07
20
05
%Δ 2014-2013 = 5%
150000
100000
%Δ 2014-2013 = 7.1%
50000
20
05
20
07
20
09
20
11
20
13
20
03
0
20
03
20
05
20
07
20
09
20
11
20
13
20
20
03
0
year
Non-GHC
GHC
Red Snapper Landings
AL
NWFL
TX
ALL
SWFL
150,000
100,000
50,000
20
03
20
05
20
07
20
09
20
11
20
13
0
150,000
100,000
%Δ 2014-2013 = -59%
50,000
%Δ 2014-2013 = 82%
13
20
11
20
20
09
07
20
05
20
03
20
13
20
11
20
09
20
07
20
05
20
20
03
0
year
Non-GHC
GHC
Number of Red Snapper Trips
AL
NWFL
TX
ALL
SWFL
3000
2000
1000
13
20
11
20
09
20
07
20
05
20
20
03
0
3000
%Δ 2014-2013 = -28%
2000
1000
%Δ 2014-2013 = 161%
13
20
11
20
09
20
07
20
05
20
03
20
07
20
09
20
11
20
13
20
05
20
20
03
0
year
Non-GHC
GHC
II. GHC vessels only:
Comparisons through time
1.
2.
3.
4.
Allocation of trips & fish over the
season
Allocation of fish per customer
Trip duration
Discards
Overall trip timing
100
50
0
# Trips
150
Weekly Trips by EFP Vessels
0
10
20
30
40
Week of year
Range 2009-2013
Median 2009-2013
2014
2013
50
Trips retaining red snapper
1000
1,295
59.1%
500
9.6%
495
31.3%
0
# Trips
1500
Cumulative Trips Retaining Red Snapper
0
10
20
30
40
Week of year
2014 season
2013
2011
2014
2012
50
Allocation of red snapper landings
Cumulative Number of Landed Red Snapper
60,000
54,907
52.1%
40,000
# Fish
11.1%
30,187
20,000
36.8%
0
0
10
20
30
40
Week of year
2014 season
2013
2011
2014
2012
50
Fish per angler – all fishing
Red snapper per angler
Δ82%
1.5
No SW Florida
1.36
1.24
Fish/Angler
1
1.09
1.01
1.06
0.99
1.05
0.99
0.81
0.77
0.68
0
.5
0.55
2003 2004 2005 2006 2007 2008 2009 2010 2011 2012 2013 2014
• These calculations include
all trips, even if no red
snapper was retained
• 2014 increases can mostly
be explained on the basis
of 82% increases in red
snapper landings
• BUT trips retaining EFP
species have increased
by 161% for red snapper
• So how are headboat
owners closing the gap?
Red snapper/angler on red snapper trips
Red Snapper per Angler (base=2013)
.2
0
-.2
-.4
-.6
Change in Fish/Angler (2013=0)
All Red Snapper Trips
2003 2004 2005 2006 2007 2008 2009 2010 2011 2012 2013 2014
Year
Estimates from regression model with vessel and month fixed effects. Weighted by anglers per trip
All landed fish/angler
All Landings per Angler (base=2013)
1
0
-1
-2
Change in Fish/Angler (2013=0)
2
All EFP Species Trips
2003 2004 2005 2006 2007 2008 2009 2010 2011 2012 2013 2014
Year
Estimates from regression model with vessel and month fixed effects. Weighted by anglers per trip
Non-EFP reef fish/angler
Non-EFP Reef Fish Landings per Angler (base=2013)
2
1
0
Change in Fish/Angler (2013=0)
3
All Red Snapper Species Trips
2003 2004 2005 2006 2007 2008 2009 2010 2011 2012 2013 2014
Year
Estimates from regression model with vessel and month fixed effects. Weighted by anglers per trip.
Reef Fish are species managed under the Reef Fish FMP
Changes in trip duration
Change in Probability of a Daytrip (base=2013)
.1
0
-.1
-.2
Change in Probability (2013=0)
.2
Red Snapper Trips Only
2003 2004 2005 2006 2007 2008 2009 2010 2011 2012 2013 2014
Year
Estimates derived from a linear probability model with vessel, month and weekend fixed effects.
Red snapper discard rates
Relative Discard Rate of Red Snapper (Discard/Catch)
1
.4
.6
.8
Δ-36%
.2
Relative Discard Rate (2013=1)
1.2
Trips with positive red snapper catch
2004 2005 2006 2007 2008 2009 2010 2011 2012 2013 2014
Year
Estimates derived from a Poisson regression model with vessel fixed effects and total red snapper
catch as an exposure variable.
Red snapper CPUE
Relative Catch Rate of Red Snapper (Catch/Angler Hour)
1.2
1.4
Relative CPUE (2013=1)
1.6
1.8
Trips with positive red snapper catch
.8
1
Δ-11%
2004 2005 2006 2007 2008 2009 2010 2011 2012 2013 2014
Year
Estimates derived from a Poisson regression model with vessel fixed effects and estimated
angler-hours as an exposure variable.
Red snapper
(discards/angler-hours)
𝐷𝐷𝐷𝐷𝐷𝐷𝐷𝐷𝐷𝐷𝐷𝐷𝐷𝐷𝐷𝐷
𝐷𝐷𝐷𝐷𝐷𝐷𝐷𝐷𝐷𝐷𝐷𝐷𝐷𝐷𝐷𝐷
𝐶𝐶𝐶𝐶𝐶𝐶𝐶𝐶𝐶
=
𝑋𝑋
𝐴𝐴𝐴𝐴𝐴𝐴𝐴𝐴𝐴𝐴𝐴𝐴 − ℎ𝑜𝑜𝑜𝑜𝑜𝑜𝑜𝑜
𝐶𝐶𝐶𝐶𝐶𝐶𝐶𝐶𝐶
𝐴𝐴𝐴𝐴𝐴𝐴𝐴𝐴𝐴𝐴𝐴𝐴 − ℎ𝑜𝑜𝑜𝑜𝑜𝑜𝑜𝑜
57% of 2013
levels
64% of 2013
levels
89% of 2013
levels
Revenues/profits?
• An ongoing topic of research
o Cost/revenue survey for 2014 (ongoing)
• Little obvious evidence of overall increases
in customer demand (i.e. more anglers/trips)
• BUT
1.
Out-of-season trips are now of a higher potential quality
to anglers and could command a higher price
2. The shift toward more day trips may enhance net
revenues
•
Higher fees with less than proportional increases in fuel
costs
Conclusions
There is preliminary evidence that,
relative to recent seasons, the EFP
has:
1.
2.
3.
4.
5.
Dramatically spread out the
allocation of EFP species across the
season
Provided many more anglers with
the opportunity to fish for EFP
species
Reduced discards of EFP species
Allocated landings of EFP species
over a broader population of
anglers (due to lower landings per
angler)
Increased the share of non-EFP reef
fish species in landings relative to in
previous derby seasons
Ongoing research
• More rigorous treatment of the without-EFP
counterfactual for GHC boats
o Quasi-experimental vs. time series approaches?
• Changes in trip-taking behavior with respect
to weather?
• Pre-test of online angler survey using 2014
data
o Working with Vic Adamowicz and Patrick Lloyd-Smith
o Estimation of consumer surplus from EFP and possible extensions
o 2015 season data currently being gathered for the final survey
Thank you