Not much, at face value. One could even claim that lotteries are quite antithetical to the spirit of savings – how could one, in good conscience, compare potentially reckless and addictive gambling with the perseverance and self-discipline that savings demands?

Funny thing is, it turns out that in some cases, they are not very different mathematically at all.

#### Haitian borlettes

NY Lottery Numbers Displayed Outside Borlette (Courtesy: Kim Wilson)

I recently read this fascinating paper titled Savings and Chance: Inclusive Finance and The Haitian Lottery on the Haitian lottery institution surrounding borlettes. Participants bet on the numbers drawn in U.S. state lotteries at kiosks, and payouts are made based on some combination of two to five digits.

Operators add their twists to this, but here’s how one of the simplest forms of this works – choose three numbers between 1 and 100, bet \$1 (or 1 gourde, the local currency) on each, and wait for the radio to announce the U.S. state lottery numbers. If your first number corresponds to the lottery, you get a \$50 payout. If your second number corresponds, you get a \$20 payout. And if its your third number, its a \$10 payout.

On any given day, for every \$1 coming in, the expected payout for a kiosk is therefore (\$1 x 0.5 + \$1 x 0.2 + \$1 x 0.1) = \$0.80. This is also the expected return for folks who play the borlette over a long enough period of time (and many do – they play a small amount with a high degree of regularity). That’s a -20% return, on average.

In what universe does a -20% return seem like a good ROI on savings, you may ask – mathematically, at that. Surely folks would save better if they simply saved under their mattress?

#### Paying to Save

This is where the caveat, “some cases,” comes in. Sure, the plain vanilla ROSCAs where n members save \$m per meeting and hand \$(n x m) to one member each meeting has a 0% return, ignoring time value of money. And savings groups that have the luxury of depositing their funds in a bank will actually make a positive return.

Many other savings setups that are common do have a cost element though. Consider the following two types that are widespread:

Organizer takes one payout: Groups often need a promoter who shepherds a complete payout cycle or two. As remuneration, the promoter takes one payout, often the first. Thus, if there are n members, the savers will save for (n+1) cycle. The return in these cases will be -100%/(n+1).

This example  from a Women’s World Banking Report is a bit dated, but the general thrust holds true across regions of the world where savings groups are formed via promoters:

On February 26th, 2003, Bethania finished a ROSCA which had five participants, each of whom contributed RD\$100 for 60 days, equivalent to RD\$6,000 each or RD\$30,000 for the entire ROSCA. The payout was RD\$5,000 every ten days and the pay out sequence was determined by lot. Bethania, as the ROSCA organizer, was entitled to the first payout, so she was able to gather this lump sum just ten days after she had organized the ROSCA. She received this without contributing any money to the ROSCA. This was her fee for organizing and managing it.

Bidding ROSCAs: In these ROSCAs, members submit sealed bids for the right to receive money in every meeting. This effectively serves as an interest payment on the savings of others, since the member gets the pot minus his or her bid amount. There are various ways of running this, and here is one example outlined in a recent paper by Tanaka and Nguyen that looks at Vietnam:

A winning bid turns into a discount to the other bidders who have not received the pool. In each meeting, the one who submits the highest sealed bid wins the pot, and the members who have not won the pool pay the full fixed amount minus the winning bid. Those who win the pot in earlier meetings get no discount, thus contribute the full amounts. The winner receives the pot, and pays a commission to the host. The cycle ends when the last member receives the pool. The winning bid of the last receiver is zero since he/she is the only bidder. Thus, the last member receives the full amount of contribution from each of other members.

Since members in such bidding ROSCAs determine their own price for the pot, and it varies from round to round, the cost varies quite a bit. In one of the examples cited in the paper above, they found that “the daily interest rates of the first receivers in these ROSCA are 0.90%, 0.88%, 0.56%, 0.17% and 0.10%, respectively” (p.g. 6).

There are other examples of where people will pay a premium to save. Yes, it denotes a negative return on investment, but it is still better than no return at all. Lest we forget, saving is hard, specially when one is talking about small amounts of income that is often uncertain, or irregular.

#### So how are they similar again?

The borlettes essentially function like a ROSCA with multiple payouts that occur in a random sequence, where the organizer charges a fee.

The borlettes also allow participants to mobilize small amounts of funds into a transformational amount of 50x.

One crucial element here is the frequency of payouts. If these functioned as the NY Lottery that they draw their numbers from, where the player has a bat’s chance in hell of getting a payout, this would not work. Participants actually count on these payouts to undertake costly projects, such as home repairs.

And finally, borlettes re-direct funds away from under the magic mattresses which often simply make savings … disappear.

#### “Micro-lotteries” ahoy!

Well, not exactly – it is somewhat unlikely that “micro-lotteries” will follow in the footsteps of micro-credit, micro-savings, micro-insurance, micro-mortgages etc. and be transplanted to other countries and settings. Borlettes are a very Haitian institution, and are a unique product of the need for a way to mobilize funds where there are very few options, disenchantment with savings schemes which turned out to be just that – schemes, and the juxtaposition with dreams and aspirations. (If that last bit seems like a non sequitur, check out the paper – its very relevant.)

Nevertheless, these and other practices arise from the desire of individuals and communities to put aside small amounts of money at regular intervals to receive a lump sum payout at a future date – a service for which they are willing to pay a premium. Borlettes make for a fascinating case study of a locally-relevant, highly scalable response to that desire.

Wo m e n ’ s Wo r l d B a n k i n g
C L I E N T C A S E S T U D I E S
F r o m t h e D o m i n i c a n R e p u b l i c