Annette LoVoi Appleseed Edgeworth Economics Subject: Economic Impact Model Summary Date: August 1, 2013

Transcription

1 th Street, NW 8 th Floor Washington, DC Memorandum To: Annette LoVoi Appleseed From: Edgeworth Economics Subject: Economic Impact Model Summary Date: August 1, 2013 Edgeworth Economics has developed a model to estimate the impact of a hypothetical change in the cost of sending remittances on the total volume of remittances sent from the United States. This model uses data published by the World Bank and can be updated as necessary to keep the estimates consistent with the latest available data. The model has three key inputs: price of sending remittances abroad, demand elasticity, and the volume of remittances sent. 1 The sections below outline how each component is obtained and used in the final calculation. I. Model Inputs A. Price Data on the cost of sending money abroad comes from the World Bank s Remittance Prices Worldwide database. 2 While we note that the database has limitations, 3 it is the most current and complete source of data we identified on the cost of remitting funds. The World Bank s measure of remittance prices includes both the transaction fee (i.e., the upfront cost paid by the sender at the location of the initial transaction) and exchange rate margin (a measure of how good a deal the sender is getting on the exchange rate 1 The model does not control for changes in factors other than price that may affect remittance markets. 2 The database calculates how much consumers must pay, on average, to send money from one country to another. These binational remittance flows are referred to as country corridors. Currently, the database covers 220 such corridors worldwide and is available at 3 For example, [c]orridor averages are unweighted and do not reflect the market shares of the different firms that compose the average. This may impact the calculation if some firm(s) are significantly larger than others, in which case taking a straight average may not accurately reflect the aggregate rates paid in a given corridor. Additionally, [i]n some corridors, exchange rate information was not available, and therefore prices in those corridors may be higher than the amount shown. (See

2 offered by a firm). 4 This dataset is published twice a year and is an aggregation of rates for different firms and products, 5 as gathered by World Bank staff though surveys. 6 The database publishes separate prices for sending amounts of $200 and $500. Because fees for sending larger amounts are smaller (in percentage terms), the model uses rates for the $500 transfer for the Low Base Fee scenario and the rates for the $200 for the High Base Fee scenario. The Average Base Fee scenario takes the average of the high and the low rates. The Base Fee scenario can be toggled in the Assumptions section on the model interface. B. Cost elasticity Elasticity of demand is an economic term used to describe the relationship between the price of a good and the demand for that good. Demand is considered elastic when a small change in price has a large impact on demand, and not elastic when a price change does not significantly affect demand. This is a crucial component of the model, as it quantifies the effect of the hypothetical changes to rates on the amounts sent abroad and the fees paid by migrants. Because of the lack of robust research on cost elasticities in the market for remittances, we have made a number of assumptions with respect to the sensitivity of senders with respect to cost. Using information from the World Bank on the share of GDP from remittances, we grouped countries into Low, Medium, and High elasticities. This classification is based on the logic that senders from countries which rely heavily on remittances as a significant portion of domestic economic activity may not have as many alternatives for generating income and may therefore continue to send remittances even if prices increase. Conversely, people from countries that do not rely heavily on funds received from abroad may not be affected as much if the costs of sending money go up. We classify countries in our model as follows: 7 4 The exchange rate margin, or spread, is an important portion of the remittance cost which not quoted in the transfer fee. While in some instances sent funds may be paid out to the receiver in U.S. dollars, remittance transactions are more typically disbursed in local currencies. As a result, an exchange operation is required to convert from sent to received currency. The exchange rate margin is the cost that must be incurred to conduct this exchange. 5 Firms include banks and Remittance Transfer Providers ( RTP ) such as Western Union and MoneyGram. Products include cash to cash, account to account, and online services among others. 6 The World Bank indicates that in these surveys, data is collected by posing as customers and contacting individual firms within each corridor. Researchers collected data within each corridor on the same day, in order to control for fluctuations in exchange rates and other changes in fee structures. It should be noted that the collected data is intended to serve as a snapshot of a moment in time, and that pricing may vary over time. (See

3 countries where remittances comprise less than 2 percent of GDP are classified as having high elasticity and are assigned an elasticity of -0.75; countries where remittances comprise between two and ten percent of GDP are classified as having medium elasticity and are assigned an elasticity of -0.50; countries that rely most heavily on remittances (i.e., ten percent or more of GDP) are classified as having low elasticity and are assigned an elasticity of -0.25; The interpretation of these values is that if the cost of sending money abroad increases by 1 percent, high, medium, and low elasticity countries will see decreases in the volumes of remittances by 0.75, 0.50, and 0.25 percent, respectively. C. Volume of funds The base annual volumes of remitted funds, by country, are from the World Bank s Migration and Remittances Data. 8 The estimated total volume (i.e., remittances that would be sent under the hypothetical price reduction) is calculated based on actual historical volume, as well as the assumed fee reduction 9 and price elasticity. 10 Once the base and estimated volumes are calculated, the appropriate cost of remitting is applied to each scenario to calculate total fees. The differences between base and estimated volumes and total fees serve as the bases of Additional Funds Sent and Total Fee Savings, respectively. II. Model Outputs The two key model outputs are Additional Funds Sent and Total Fee Savings. The measure of additional funds sent is calculated as the difference between the base and estimated volumes of remittances. This represents the additional funds migrants can be expected to remit if the cost of transferring money goes 7 While research on the topic of remittance elasticities is sparse, at least one paper corroborates our assumptions. (See Gibson, John, David McKenzie, and Halahingano Rohorua How cost elastic are remittances? Evidence from Tongan migrants in New Zealand. Pacific Economic Bulletin, Vol. 21(1), pp ) This paper studied migrants from Tonga, which has one of the highest remittances as shares of GDP, at approximately 20 percent. The study estimated an elasticity of -0.22, which is in line with our model s assumption. 8 The World Bank publishes datasets describing remittance flows, migration trends, and sovereign ratings, in addition to publically available survey data regarding those topics, at 9 The estimated fee is calculated by subtracting the assumed fee reduction (in percentage points) from the base fee percentage. That is, if the current cost of sending money is 5 percent, and a 0.5 percent reduction is assumed, the resulting fee would be 4.5 percent. 10 The exact relationship is defined as Destimated = Dbase * [1 + (Pestimated / Pbase - 1) * ε] where D is total volume, P is price, and ε is the cost elasticity.

4 down. The fee savings represent the difference between base and estimated total fees. Because our research indicates that the demand for remittances is not highly sensitive to price, the model restricts the elasticity assumption to values between zero and negative one. This restriction implies that that the change in demand is lower than that in price, and the total amount of fees should decrease if rates go down. 11 The model presents results for each of the 28 countries for which price data is available in World Bank s Remittance Prices Worldwide database. 12 It is important to note that this is a historical model. The model uses the latest available data, but makes no assumptions about future prices or growth in the volumes of remittance flows. Figure 1 highlights the various components of the model s interface. 11 Hypothetically, if the cost of an individual transaction decreases, the total number of transactions would likely increase and the effect on total fees paid could be ambiguous. Consider, for example, a scenario where the price of a transaction is $1 and an individual chooses to send money once. If the price goes down to $0.50 and the same individual conducts two transactions, the total amount sent home may be higher, but the total fees paid would be the same. If only one transaction occurs at $0.50, however, the total amount sent may remain the same, but the total fees paid would have decreased by 50 percent. 12 These include Brazil, Cape Verde, China, Colombia, Dominican Republic, Ecuador, El Salvador, Ethiopia, Ghana, Guatemala, Guyana, Haiti, Honduras, India, Indonesia, Jamaica, Kenya, Lebanon, Liberia, Mexico, Nicaragua, Nigeria, Pakistan, Panama, Peru, Philippines, Thailand, and Vietnam.

5 FIGURE 1 MODEL INTERFACE Country toggle: Select one of 28 countries for which World Bank data is reported. World Bank data Base Fee scenario: Can be set to Low, Average, and High. Hypothetical fee reduction (must be between zero and the base fee). Calculated as Average Fee x Total Volume. Cost elasticity based on remittances as share of GDP in receiving country Calculated based on base price, base volume, elasticity, and price change. Estimated minus Base