Piloting + sacking

While Accra continues shamble along as only it can, I have found myself less captivated by its peculiar rhythms since I’ve returned. Familiarity leads to monotony; or is it the other way around? Yes—I think it’s the latter.

In the past two weeks work has been the focus. We’ve reached a major hurdle—the beginning of the pilot—and I think we’re in the process of clearing it. But this is not a simple run-and-jump deal. Every miniscule movement of our collective striding legs is an excruciating coordination of myriad twitching nerves.

But what’s the pilot anyway? For that matter, what’s the project for which it serves as preliminary research? These are questions I have not yet answered in this rambling blog, and I will try to do so in the following paragraphs.

The project I was hired to work on is an Interest Rate Sensitivity Study for individual microloans. The set-up is as follows. I work for Innovations for Poverty Action (IPA), a US-based NGO run by development economists. Most generally, our output is papers to be (hopefully) published in academic journals. More specifically, those papers tend to be write-ups of field experiments in economic development—particularly in the science of evaluating development programs. The question that guides IPA’s research is, broadly: What works in development, and why?

In some remote corner of that question lives microfinance; in some nook in that corner are microloans; in one cranny of that nook are microloans; and on one grain of sand in that nook is the individual loan product offered by Opportunity International – Sinapi Aba Savings and Loans (OISASL). IPA has partnered with OISASL to investigate the following specific questions: How do individuals at various poverty levels respond to different interest rates in microloans? Do particular interest rates attract certain types of borrowers? Are people’s abilities to repay dependent on the interest rate offered?

To answer these questions, we have designed the study that has just entered the pilot phase. For the first two months of my stay here, we pounded the hot pavement of Accra cataloguing some 7,000 businesses large and small—everything from sidewalk stands selling mangoes to more sophisticated operations in permanent buildings with multiple employees. These 7,000 were divided into 125 geographic clusters (think of a cluster as comprising all the businesses on a given city block), and to each cluster we will randomly assign one of four different interest rates.

Next we developed a poverty-level assessment (in the form of a short questionnaire) and promotional flyers inviting first-time customers with OISASL to apply for loans. Then comes the labor-intensive step—marketing. It proceeds as follows: an OISASL marketer approaches one of the catalogued businesses, asks the owner the poverty-level assessment questions (e.g. How does your household get drinking water? Do all children age 6-17 in your household attend school regularly? What kind of cooking fuel do you use?), tells him a little bit about our promotional offer, and hands him a flyer advertising loans at the interest rate we have assigned to his cluster, inviting him to come to the branch and fill out an application. Repeat 6,999 times, and we’re done with that step.

Once the business owners have flyers and we have their poverty-level information, it just remains to track their responses. (A) Who comes into the branch? (B) Who continues with the application process? (C) Who is approved? (D) Who is able to repay?

We will accumulate the data necessary to answer those four questions in the year or so following the marketing campaign. Then all that remains is analysis and, finally, writing the paper itself (happily we have Dean Karlan, a rising star in development economics, to do the bulk of the last step).

So that’s the project in a nutshell. The pilot is essentially a short-term, small-scale implementation of that mess. We have selected four clusters (one at each interest rate) and are now marketing to the businesses in them. Our goals for the pilot are (1) to refine our marketing technique, (2) to streamline the reception of study participants in the branch, and (3) to get some estimates of (A) – (C) above. Once we have a sense of what percentage of people will come to the branch, we can set about preparing to handle the additional workload associated with the full study. We will not attempt to study (D) in the pilot, as repayment data only builds up over the life of the loan (typically 6-12 months).

For (2) above I probably should have written “damage control”. It is ultimately a dangerous thing to advertise different rates to different people, especially when many of those people are Ghanaian petty traders, a notoriously gregarious variety. The concern is that we will have a lot of irate customers coming into the branch and laying into the individual loan officers: My friend owns a shop just across the street, and he got a lower rate than me. Give me the lower rate! One reason for randomizing interest rate at the cluster level (as opposed to at the individual level) in the first place was to minimize this issue. If we offer everyone in a given cluster the same interest rate, then hopefully friction will only occur at the edges where it borders on other clusters.

Incredibly, Ghana exhibits extraordinarily resistance to even this most rudimentary instance of advance planning. As I said before, many of the businesses we’ve catalogued are street-side stands. Sometimes goods are laid out on a blanket spread on the sidewalk; other times they are stacked on a table. In downtown Accra these stands are ubiquitous. They seem to occupy 75% of passable sidewalk, and in many areas, especially where there are “rough” (dirt) roads, they spill out into the street. While this makes walking in the city an exhausting exercise in dodging and jostling, it also means that one can buy anything from fresh produce to toilet paper to stereo systems without walking more than a couple blocks. From the seller’s perspective it means a wealth of loosely (if at all) regulated commercial space in a heavily-trafficked market; and a typical sidewalk stand operator has staked out a physical location and a regular customer base that he returns to everyday.

The Accra Metropolitan Assembly (roughly analogous to the Mayor’s office) has decided to put an end to this in one fell swoop. In preparation for Ghana’s Golden Jubilee 50^th Anniversary on March 6 of this year, the AMA is undertaking a number of ambitious projects to improve the look and feel of the capital. While these improvements are ultimately intended to benefit the residents of Accra, most people readily admit that the reason they’re being made—and made quickly—is so that Ghana can put her best foot forward for all the important folks (read “first world tourists and politicians”) coming in for the celebration. In addition to widening some major roads and installing trash cans in public areas (wow!), on Monday they began the systematic “sacking” of all informal street vendors, from hawkers (people selling on foot, carrying their goods in large baskets balanced on their heads) to tabletop stands.

“Sacking” really means forced relocation to one of two locations that have been set aside for the vendors. Most Ghanaians I’ve talked to are thrilled with the results—in the space of three days the downtown streets have been rendered passable again, and the whole area is undeniably cleaner and less chaotic. Until Monday afternoon Makola, a twelve-square-block area just a few minutes walk from the OISASL office, was the densest market in Accra, with vendors completely filling the streets. On Tuesday morning it was unrecognizable: wide, empty thoroughfares with cars timidly poking along, still not sure if they really belonged there. In corners and against the sides of buildings were piled great drifts of broken and abandoned wooden tables and racks—the dusty, brittle bones of the marketplace.

There is reason to believe that this new arrangement does not constitute a free lunch. For one, the largest area that the AMA has set aside for the vendors is just beside Kwame Nkrumah Circle (known here as “Circle”), which has the dubious honor of being Africa’s largest roundabout. It is easily the most congested area in the city, and the addition of thousands of hawkers and vendors—not to mention their customers, who would otherwise have no reason to go there—is sure to exacerbate the problem. Fortunately Circle is not on the tour bus route.

Above, I introduced “sacking” by way of saying that Ghana had resisted our attempts at planning ahead to avoid complaints about offering different interest rates to different people. About half of our 125 clusters are in areas that have been, or will soon be, sacked. So whereas we had hoped that those clusters, and the vendors comprising them, would remain geographically distinct (even if only separated by a street), now they will be busted up and the sidewalk vendors among them will corralled as tightly as possible into one massive pit.

Recipe for disaster? Maybe. Then again, if things were easy, they’d be too easy.

Much love and cheer from Accra,

Jake