General thoughts on Uspantán

I'm getting ready to clear on out of lovely Uspantán tomorrow for
probably the last time. This has been a really enjoyable stay, and
productive (I think), although some of statistically significant
findings that came out of ideas gathered here are really sensitive to
model specification, and other ideas I can't test. Still, I have high
hopes for these ideas about jurisdictional optimality and social bonds
and collective action problems.

Although I really get a kick out of conducting these interviews and
being able to talk to the locals, the best part of this segment has been
the fantastic set of Estadounidenses that I've bumped into here--the
three local Peace Corps volunteers (without whom, this segment of my
trip wouldn't have been nearly as successful), and the folks from
Louisville down here putting in water purification systems (without whom
I wouldn't have eaten nearly as well--though I would have eaten several
more combination hamburger/fried chicken sandwiches).

Basically, my story about Uspantán is this:

First, this is a geographically huge municipality--they tell me that it
takes about ten hours to travel from end to end, which seems about
right, based on my limited experience. In a jurisdiction of this size,
even a municipality, the statistics we use to measure local outcomes
mask a tremendous amount of variation that I wish I could get my hands
around. Statistics at the village level, in a place like this, would be
fantastic.

Even so, I've found out one thing:

The thing I've found out--Uspantán isn't the low outlier that it first
seemed to be. Upon some observation, I suspected that some of the
numbers in our survey were inaccurate, so I went in to talk to the
treasurer, and after three attempts, finally caught up with him. I
double-checked some of the strange values in our dataset (municipal
employees and percent of the local budget in forestry) and found that
the old values were too low. I entered the new values and re-plotted
the regressions, and although Uspantán still sits a little below the
regression line in most of the models I've been using, it's a lot closer
than it was before. This means that Krister's institutional approach
explains Uspantán better than it seemed at first.

And I've come up with some good ideas which I might be able to test:
1. Social bonds severed by the civil war may help communities surmount
collective action problems to manage forests in a more sustainable way,
and to avoid tragedy-of-the-commons situations.
2. Too-large jurisdictional boundaries can also create a
tragecy-of-the-commons, where municipalities face pressures from
individuals who favor too-restrictive and unrestrictive forestry policy,
and where individuals feel compelled to harvest trees to quickly, lest
their neighbors do it for them. Smaller jurisdictional sizes with de
facto powers of enforcement solve this problem, by creating clear
community "property rights" which encourage sustainable practices (since
the community can ensure that the trees won't be cut down by others,
they have more reasons to harvest them gradually, in a sustainable way)
3. Social bonds also seem to matter in helping communities avoid
internal tragedies-of-the-commons.
4. And although the part of me that's been socialized as a rational
choicer wants to pooh-pooh this, constructivist-like ideas about people
developing an environmental consciousness have come up again and again
interviews, and seem to be supported by the statistical significance of
variables that measure education in several models. I just wish I had a
more convincing way to get at this, and to avoid the reverse-causality
associated with these kinds of models (of course people who have, for
example, an economic interest in sustainable forestry will be keen to
talk about conservation and their development of an environmental
consciousness).

Laj Chimel

Laj Chimel means "Little Chimel," because it's a village built on the
site of the former village of Chimel. It's little because 14 families
live there, but 70 families lived in old Chimel.

Chimel is the home of Rigoberta Menchú, who won the Nobel Peace Prize
back in the early '90s for helping to publicize the plight of rural,
indigenous Guatemalans in the Guatemalan civil war. Although her book,
_I, Rigoberta Menchú_ has been criticized (not least for factual
inaccuracies), it's a pretty accurate story if taken as a typical
account of Guatemalans during the armed conflict here.

One of the local peace corps volunteers and one of the municipal
employees here had some work to do up there--they're promoting an
ecotourism project up in the village--and she agreed to take me and
several other gringos--a bunch of folks from Louisville, KY working on
water purification projects down here--up to the village to check it
out, and so I could talk to the people up there about their efforts to
conserve local forests.

The trip was incredible.

Not only did we get to take a hike through the beautiful and incredibly
rich cloud forest up there, but the locals had put together this
incredible rope swing which we were able to try out, we got a look at a
local lagoon which fills up with water in the dry season and dries out
when it rains (good for swimming--I'll put the pictures up when I can
get a wireless connection) we saw some beautiful scenery, I was able to
conduct an interview, and we good to hear the incredible and painful
story of one of the local women's experience during the civil war,
including accounts of being raped, her baby daughter dying, watching
friends tortured and killed, and watching her village be burned to the
ground.

A little background:

Supported by the US, the Guatemalan military regime fought against
Marxist guerrillas (themselves largely a product of earlier, arrogant
and misguided US interventions in Guate.) using techniques borrowed from
US efforts against the Vietnamese. Literally, the Guatemalan Army did
its best to clear the countryside of peasants, so there was nobody who
could support the guerrillas. Mostly, the peasants just wanted to be
left alone, but they got caught in the middle, with terrible results.

But now, they're back, trying to rebuild (and making a good go of it, by
the looks of it.) After buying back their land from large landowners
who had taken it over during the war, two of the original families have
returned, with twelve other families. Beautiful place, though I can't
imagine returning home to a place where you've experienced so much pain
and heartbreak...

Negative binomial model

Went back and did a little reading on the poisson model--sad to say, my
model (below) doesn't pass the post-estimation tests. So I tried a
negative binomial model (thanks to Ying Lu for teaching us all this
stuff), but I get two different sets of results, depending on how I set
the dispersion. Think I'll need to wait to get back home to reread my
notes on this stuff.

Poisson Regression

Look at these beautiful z-values!

I've been fighting with a badly skewed dependent variable--the
percentage of the municipal budget in forestry. It occurred to me that
the distribution resembles a poisson distribution. But I'm not sure
that this is right. Poisson regression is for count variables--number
of people who fall out of the back of Guatemalan pickup trucks, for
example. But this is a proportion... I don't think I have straight
budgetary size, which would be better...

This doesn't work as well with a personnel variable--the number of
people working in forestry.

Iteration 0: log pseudolikelihood = -164.60041
Iteration 1: log pseudolikelihood = -158.84706
Iteration 2: log pseudolikelihood = -158.82052
Iteration 3: log pseudolikelihood = -158.82052

Poisson regression Number of obs
= 50
Wald chi2(12) =
175.07
Prob > chi2 =
0.0000
Log pseudolikelihood = -158.82052 Pseudo R2 =
0.4930

------------------------------------------------------------------------------
| Robust
for_budge~08 | Coef. Std. Err. z P>|z| [95% Conf.
Interval]
-------------+----------------------------------------------------------------
NGO_demands | .5640684 .189278 2.98 0.003 .1930903
.9350465
NGO_ec_suppo | -.3174905 .1405181 -2.26 0.024 -.5929009
-.0420801
rel_cit_dema | .6391908 .2348854 2.72 0.007 .1788238
1.099558
import_to_ci | -.3157467 .2260987 -1.40 0.163 -.7588921
.1273987
import_cent_ | .1148308 .1097506 1.05 0.295 -.1002765
.3299381
user_et~g_08 | .5093573 .3601709 1.41 0.157 -.1965647
1.215279
elite_educ | .0838478 .0705506 1.19 0.235 -.0544288
.2221245
pop_dens_01 | .0018278 .000772 2.37 0.018 .0003147
.003341
size | .000385 .0001025 3.76 0.000 .0001841
.000586
empl_per_~08 | 2.447348 3.064496 0.80 0.425 -3.558954
8.453651
pct_for_c~08 | .0180918 .0111026 1.63 0.103 -.0036688
.0398525
import_~v_08 | .4119408 .1828592 2.25 0.024 .0535434
.7703382
_cons | -1.051291 1.179759 -0.89 0.373 -3.363577
1.260995
------------------------------------------------------------------------------

Why...

...are Guatemalan pillows either lumpy as all get-out, or hard as a soft
rock?

Jurisdictional Flexibility

Elsewhere, I've spent some time discussing my theory of "jurisdictional
optimality," in which larger municipalities are more likely to be
further from the optimal jurisdictional size. Optimality, in this case,
is defined (following from Ferejohn and Weingast 1997) as jurisdictional
boundaries being appropriately shaped and sized so that the benefits of
a given policy are shared by the same group of people who share the
costs of that policy. In more concrete terms, in this case, this means
that jurisdictional boundaries are drawn so that the costs of enforcing
sustainable forestry regulations are borne by the same group of people
who will experience the benefits.

If the hypotheses regarding jurisdictional optimality tested here are
accurate (as statistical evidence seems to indicate), there are a few
implications that bear directly on the structure of decentralization
reforms in general, and Guatemala's process of decentralization in
particular.

First, because different policy tasks will affect differently-sized and
differently-shaped geographic groupings of people, it is important for
decentralization reforms to incorporate mechanisms by which
jurisdictions can be flexibly bounded—either smaller or larger,
depending on whether costs and benefits are distributed optimally among
larger or smaller or differently shaped jurisdictions.

Second, it is important that these jurisdictional sizes and shapes are
allowed to change with relative ease—because it is unlikely that
jurisdictional sizes and shapes will be correctly anticipated by
reformers when decentralization reforms are initially implemented, and
it is equally unlikely that jurisdictional needs will remain the same
over long periods of time, as population distributions, ecological
dynamics, and technology (among other factors) change.

In the case of Guatemalan decentralization reforms in the forestry
sector, some of the elements needed for optimum jurisdictional
flexibility are present, while other elements are missing.

It is relatively easy, for example, for Guatemalan municipalities to
form mancomunidades, or co-management regimes, in which several
different municipalities can join together in a special-purpose
jurisdiction. Some examples are the jurisdiction created by the rural
municipalities in the environs of Quetzaltenango for development and the
attraction of NGO, IGO, and aid organization funding, and the
co-management regime under construction between Cantel and its neighbors
for the environmental protection of the Samala river. Where
municipalities are too small, it is easy to create larger jurisdictions
to approximate jurisdictional optimality.

What is largely absent from the Guatemalan decentralization scheme,
however, is a legal way to constitute smaller-sized jurisdictions, and
combinations of smaller sized jurisdictions, perhaps at the level of the
village. As a result, where jurisdictions are small (in small
municipalities like Santa Catarina Barahona and Zunil) it is easier to
build larger units for optimal governance. Where jurisdictions are
large, however, it is difficult to create effective, lower-level,
smaller units of governance for policy tasks which are best managed in
units of smaller geographic size. Forestry and watershed management, for
example, might be policy tasks best managed at the level of relatively
smaller jurisdictional units such as villages (Aldeas) inside of
municipalities.

While Aldeas do have legally recognized community organizations (COMUDES
or municipal councils for development), these organizations primarily
serve purposes of interest aggregation and articulation, passing
requests on to the Mayor and municipal council, not playing any role in
self-governance.

Guatemalan Patent Medicine

When I was in the Totonicapan market this morning, I heard a couple of
gentleman shouting the benefits of their "natural remedies." I got a
kick out of the fact that one such product was being advertised as good
for "headaches, colic, anger, and fear..." among other things. It
doesn't take too many visits to New England county fairs (or the
Rockland, Maine Lobsterfest) to know that this sort of thing is alive
and well in the states, as well.

I've been gettin' some hard travelin'

...I thought you know'd...

Way down yonder in Totonicapan, way down the road
In the back of a picop, dusty 'n rough
Couple of Guatemalans helped me find my bus

I've been gettin' some hard travelin' Lord

That may be a little too obscure of a reference, but it's been bouncing
around in my head for a couple miles. Or kilometers, as they use them
down here more often.

Spent the night in Totonicapan last night--wanted to at least poke my
head into the municipality which has such unusual forestry-related
outcomes (very high) and hang around for the enormous Saturday market,
which seems to take up about half the city (though the streets look so
different with market stalls in them, I have no way of knowing exactly
how far the market extends--I found myself repeatedly turned around when
distracted by beautiful looking carrots or really cheap kitchen knives.

Incidentally, a very high quality (though rather utilitarian) carving
knife down here goes for about $3.50. That's more than Bolivia, where
they were about a dollar, but still a hell of a deal. I imagine they're
cheaper in Brazil, as well, as they're all made by Tramontina, the very
high-quality Brazilian cutlery firm.

I was hoping that once I got to Totonicapan, I would discover direct
buses from there to Quiché, where I would then be able to move on easily
to Uspantán, but I had no such luck. When I inquired about pickup
trucks, the answer was in the affirmative (although a hesitant
affirmative, as I am only a stupid Gringo).

But I decided that I would try to catch a pickup truck for about an
hour, and if I didn't have any luck, I would walk down to the bus
terminal and go that way.

But I thought the direct route from "Toto" to Quiché is much shorter,
and I thought it would be a lot less complicated, though I'm not sure I
was right about either of those things. It is shorter in distance, but
I'm not sure how much time I saved. Nevertheless, it was a hell of a
beautiful (if dusty) ride. Definitely worth the extra effort (and
embarrassment).

So I spent about two hours wandering through the market this morning,
and I got some good pictures and video, I think. Then I headed back to
pack up, and I was out of the hotel room at about 11:30.

I found the correct corner for the Quiché pickup trucks, and while
there, I was approached by a friendly and helpful Guatemalan guy with
his daughter, who were headed my way, and pointed me in the right
direction. We found a truck, hopped on board (hopefully without bumping
too many heads along the way) and we were off.

The first portion of the trip was incredible. First, through the
beautifully maintained forests of Totonicapan municipality (one of the
only places in the country without a deforestation problem) and then,
out into beautiful pastoral landscapes (where the forest has been cut
for agricultural purposes). It's hard, though, looking at those farms,
to not think about the way we (the US) more or less unthinkingly plowed
the great plans under and completely changed the ecosystem of much of
the landscape between the Appalachians and the Rockies.

Stunning country. As I told someone, like a sublime combination of
Virginia, Pennsylvania, and the Rockies.

In any event, we made it into San Antonio, where we changed pickups for
the remaining 20 minute ride into Quiché. Along the way, the Guatemalan
guy got off, but not before assuring me that the truck would take me to
the central plaza in Quiché and recommending a spot for lunch (Pollo
Campero, of course!) From there, on to a chicken bus, and on to
Uspantán. On the chicken bus, I ran into another Estadounidnense with
whom I had a nice conversation about her peace corps experience in
Lesotho. I learned a lot. For starters, it's pronounced Le-soo-thoo.

And now I'm back in Uspantán, just two rooms down from where I was a
week and a half ago. Not quite home, but it feels good to be sitting
still again.

Totonicapan

Here in "Toto" (as the guys on the bus say it) they do a couple of
things wrong:
1. Not a single diet coke or diet pepsi in town.
2. Terrible, off-key singing from the evangelical church down the road.
3. No direct buses from here to Sta. Cruz del Quiché

However, they really have a couple things right:
1. Breakfast for lunch.
2. Best "American" breakfast I've had to date--all the right elements.
Includes two pancakes, real maple syrup, ham and eggs, coffee, and corn
tortillas.
3. I think I'm the only gringo in town
4. Direct pickup truck rides to Sta. Cruz del Quiché
5. Guy greeted me on the street by saying, ¡Que bigotes!

¡Tiky Piña!

I am thrilled to announce that I have discovered the Guatemalan version
of my Peruvian beverage of choice, Inka Kola--a yellow,
bubblegum-flavored, excessively-sweetened soft drink which I pine for
while not in Peru (and which I gorge on when I am there).

There is a soda here which is identical in every way, but the packaging
is different--supposedly, it's pineapple flavored, but they're not
fooling me--I know that it's really Inka Kola flavored.