A framework for examining flows from migrant workers, with a focus on Egypt’s experience in the 1980s
One of the striking and immediate aftershocks of the recent Middle East war was the sudden stoppage of flows of workers’ remittances from Iraq and Kuwait. The effects of this were disastrous on the economies of many countries in the region (and beyond), as it became evident that the size and importance of these remittances had been grossly underestimated for many labor-exporting countries. A major factor was the revelation by thousands of fleeing workers in the aftermath of the Iraqi invasion of Kuwait that they had kept large savings in Iraq and Kuwait that were never remitted to their home countries. Among the seriously affected countries, Egypt was estimated to have lost about $2 billion in remittances during 1990 plus another $13 billion said to have been the accumulated savings of Egyptian workers in Kuwait. These unremitted savings in Kuwait alone were equal to 46 percent of the officially recorded remittances to Egypt during 1980–89. Had adequate policies been in place, a large part of these savings may well have been transferred to Egypt, providing it with much-needed foreign exchange.
This article, based on a study covering remittances over 1974–89 (see box), examines the Egyptian case against the backdrop of the growth of remittances to countries in the Middle East and a theoretical framework for analysis of such flows. It then draws policy lessons from this experience.
The growth of remittances
In the aftermath of the 1973 oil crisis, labor from different countries in the region, as well as South and East Asia, migrated in great numbers to the oil-exporting countries of the Middle East. The number of Pakistani workers, for example, jumped from roughly 500,000 in 1975 to 1.25 million in 1979. With this massive movement of labor came a movement of capital in the form of workers’ remittances to their home countries. These became the major source of foreign currency and, therefore, an important element in the balance of payments of many labor-exporting countries. Such flows, captured in the national accounts under the heading of unrequited transfers, continued to grow dramatically up to the mid-1980s in most cases, particularly relative to the flows of exports of goods and services from the home countries of migrant workers.
There was a general downward trend in the period 1985–89, but in many countries in the region the volume of remittances remained high relative to exports of goods and services. In 1989 this ratio was: 41 percent for Egypt, 24 percent for Jordan, 37 percent for Pakistan, 50 percent for Sudan, and 37 percent for the united Yemen. As a percentage of GDP, remittances remained at roughly the same level in Pakistan for 1988 and 1989 (5.6 and 5.9 percent respectively), and at a relatively high 17 percent for Egypt in 1989.
The downward trend of remittances since the mid-1980s has generally been explained by the recession during that decade following the collapse of oil prices, triggering a return of workers to their home countries. It is estimated that over 1985–89 there has been an average net outflow of 100,000 Egyptian workers each year from oil-exporting countries, including Jordan, yet Egyptian officially recorded remittances continued to rise, reaching a record $4.2 billion in 1989. Such movements in recorded remittances can be studied within a framework that classifies the different types and factors affecting them.
The longer paper, on which this article is based, entitled “Workers’Remittances as a Source of Foreign Earnings: A Study in the Determinants of Recorded Remittances in Egypt,” was written when the author was an intern at the IMF.
Understanding the flows
A simple taxonomy of remittances yields four types:
Potential remittances are the savings available to the migrant once all his expenses are met in the host economy. These represent the maximum amount a migrant can remit.
Fixed remittances are the minimum amount a migrant sends to satisfy his family’s basic needs. These can be sent either through official channels (money orders, banks, checks, etc.), in which case they are classified as fixed recorded remittances in the balance of payments accounts, or through unofficial remittances, classified as fixed unrecorded remittances.
Discretionary remittances are what the worker remits over and above the fixed amount sent either through official or unofficial channels.
Saved remittances, or retained savings, are the amount not remitted. They are represented by the difference between total savings and actual remittances in that period. Although saved remittances are a “flow” they are also a “stock” in the parlance of national accounts, since the worker can accumulate and remit them at any time, thus increasing actual remittances in any period.
For the purpose of analysis, therefore, three categories can be examined: potential remittances or total savings, comprising actual remittances (official and unofficial) plus retained savings, recorded remittances, comprising fixed and discretionary remittances, and unrecorded remittances, comprising the fixed and discretionary parts sent through unofficial channels. By understanding the dynamics behind the flow of recorded remittances, one can gain an insight into the remittance behavior of workers.
The process behind remittances
The amount the worker initially sends to his family (fixed remittances) will primarily be determined by factors such as the size of the household and the contractual arrangements with the family (whether implicit or explicit), among other factors. Whether the worker sends this amount through the official or the unofficial market will depend on the difference between the official exchange rate and the parallel (or black market) rate, as well as the cost of going through the unofficial market. This cost involves the search for a means of sending the remittances as well as the worker’s perception of the risk in using unofficial channels. The worker has no incentive to send his money through official channels if the difference between the official and parallel exchange rate (the parallel market premium) is greater than the cost of going to the parallel market. In this arbitrage process, no variables such as interest rates apply.
On the other hand, the flow of discretionary remittances is determined primarily by the difference between the real domestic interest rate and the real foreign interest rate. For the worker to remit part of his savings over and above the fixed amount mentioned earlier, the real interest rate in his home country must be greater than interest rates in the host country (or any other market accessible to the worker). Furthermore, for these remittances to flow through official channels the exchange rate difference must be greater than the cost of going to the parallel market. If either one of these two conditions is not met, discretionary remittances will not be sent through official channels. In that case, these remittances either will be sent through the unofficial market or they will remain in the host country as saved remittances.
It is possible under these conditions for recorded remittances to fall during a period, even when actual remittances are on the rise. This would happen when the exchange rate difference increases, causing a shift from the use of official channels to unofficial channels. Actual remittances may, on the other hand, increase for several reasons, including an increase in total savings due to a greater number of workers, interest rate differences in favor of the domestic market, and an increase in the flow of previously saved remittances. In theory, recorded remittances could also increase while actual remittances are falling. This would be the case if policy changes induce a shift from the unofficial to the official market, increasing the official flow of remittances. On the other hand, actual remittances may decrease, due, for example, to the fall in total saving resulting from a smaller number of workers. This is probably what occurred in Egypt between 1986 and 1989, when recorded remittances increased, while actual remittances were probably decreasing.
When a parallel market exists and the government wishes to increase the flow of recorded remittances, it could devalue its exchange rate. A devaluation will reduce the difference between the parallel and official rates and will make it attractive for workers to remit through official channels, irrespective of the interest rate structure. This will increase the flow of official remittances, other things remaining constant.
A government could also attract the savings that would otherwise have remained in the host country by sufficiently increasing domestic interest rates relative to those in the host country of the workers. However, this increase does not necessarily mean that there will be a rise in the official flow of remittances if there still exists a strong incentive to remit through the unofficial market. Stringent policies and higher penalties for those caught operating in the black market may deter some workers from sending their money through unofficial channels. This, however, will not affect the flow of discretionary remittances, but only the flow of fixed recorded remittances. Finally, political instability in the home country does not appear to affect the flow of fixed remittances but will affect the flow of discretionary remittances.
Empirical results from Egypt
The preceding theoretical framework was applied to a study of remittances to Egypt during the 1980s. This study showed that the flow of recorded remittances is quite sensitive to exchange rate differences between the official and parallel market. The estimated elasticity of the ratio of the parallel rate to the official rate, varied from a low of 0.9 to a high of 1.18. Thus a 10 percent reduction in the premium given by the parallel market would increase the flow of official remittances between 9 and 11.8 percent.
Following the 1987 devaluation of the Egyptian pound by 37 percent, recorded remittances increased from $2,515 million in 1986 to $3,604 million in 1987, a 43.3 percent increase (see chart). The part remitted in cash (as opposed to transfers of goods) jumped from $464 million to $1,446 million, confirming the high sensitivity of recorded remittances to exchange rate differentials.
Although it was not possible to test the significance of the interest rate differentials on the flow of actual remittances, two observations point to the importance of interest rates as a determinant of the flow of actual remittances. Islamic companies established in Egypt in the early 1980s were offering returns on deposits of approximately 24 percent in nominal terms compared to 13.25 percent on ten-year Egyptian government bonds. It is estimated that by 1987 these companies had accumulated deposits of $9 billion from workers abroad, tempted by the high (but still negative real) domestic interest rate. These remittances, for the most part, went through unofficial channels. Although the Government attempted to develop its own system of Islamic banking, the rates offered were similar to the traditional deposit rates and the Government failed to attract savings of workers abroad. The second observation concerns the disclosure by Kuwaiti authorities in the aftermath of the Iraqi invasion of Kuwait, of significant savings of Egyptian workers. This phenomenon points to the workers’ awareness of interest rate differentials, and its effects on workers’ decisions to arbitrage between the domestic rates and the rates in the host country.
Estimates of total workers’ savings
What was the size of the savings that workers did not remit to Egypt during the 1980s? The estimates for potential remittances, or total savings, were calculated for two years, 1982 and 1989, on the basis of estimates of the total number of Egyptian workers and their wages in the eight labor-importing countries of the region. In 1982, total savings amounted to some $7,200 million while in 1989 they had dropped to $5,400 million. The decrease may have resulted from the fall, both in the number of workers and the weighted average nominal wage, for manual and nonmanual labor. Indeed, over 1985–89 recorded remittances rose. Domestic policies such as the 1987 devaluation, the stringent policies and high penalties for using the parallel market, as well as a drawing on past savings may have contributed to the increase.
Official workers’ remittances to Egypt, 1974–89
Source: International Financial Statistics, IMF, 1989, and January 1991.
1Of exports of goods and services.
Using these total savings figures, our study recalculated the foregone earnings for the Egyptian economy as a proportion of total exports, assuming that the foregone foreign resources either escaped official channels or simply remained abroad because of unattractive domestic interest rates. Total recorded remittances for 1982 were 34.3 percent of Egyptian exports of goods and services. Potential remittances however, were about equal to the total value of exports of goods and services for that year. In 1989, the ratio was 41 percent for recorded remittances and 52 percent for total potential remittances. The difference between potential remittances and recorded remittances was equal to about 135 percent of the current account deficit of Egypt in 1982 and 68 percent in 1989. The decrease is mainly a result of an increase in the size of the current account deficit in the past years as well as a fall in total saving. If the amount equal to the difference between recorded and potential remittances—that is the foregone official earnings for the Egyptian economy—had been remitted, it would have produced a surplus on the current account between 1980 and 1986. Note that this analysis does not take into account the use of remittances by the receivers which can adversely affect the balance of payments (e.g., by consuming imported goods). This “leakage” has been shown not to be significant. Note also, that balance of payments data distinguish between in-kind and cash remittances. In that sense, consumption of imported goods through remittances is largely accounted for.
Exchange rate and interest rate differential. Our empirical evidence suggests that a policy based on the exchange rate difference can only attract to official channels fixed remittances. So long as domestic interest rates are not competitive relative to foreign interest rates, there is no immediate incentive for the worker to remit the discretionary part of remittances. In the extreme case, the worker will not remit through official channels so long as real domestic rates are less than the foreign rates. In practice this is not necessarily the case since there are other factors (discussed below) that affect the workers’ decision to remit. Thus, if there is both a reduction in the parallel market premium and an increase in the domestic interest rate, remittances—fixed as well as discretionary—will be channeled through the official market.
Other factors. Political instability in the home country, and a number of other factors, adversely affect the flow of recorded remittances. Inconsistent government policies, such as a temporary ban on imported goods, can reduce the demand for foreign currency, thus cutting the parallel market premium. If such government actions are perceived as temporary by the worker, even a reduction in the exchange rate difference will probably have no effect on his decision to remit through official channels. Thus it is suggested that the greater the variance in the government’s policies the less will be the migrant’s willingness to use official channels.
Another important factor affecting remittances is the availability of financial intermediation. Many workers use the parallel market for lack of a more efficient means of transfer. The absence of financial services in rural Egypt, for example, means that services offered by private agents are welcomed by workers. Studies show that 53 percent of migrants from rural areas used friends, relations, or both, as a means of transferring their remittances to Egypt, because of the absence of official channels.
In summary, our study indicates that, first, recorded remittances are a misleading indicator of the size of remittances relative to other sources of foreign exchange for the domestic economy. Second, to analyze remittances, a proper framework must be developed that clearly defines the different types of remittances. Third, evidence from Egypt during the 1980s suggests that exchange and interest rate differentials are important determinants of the flow of remittances. Finally, the estimated size of potential remittances (or total savings) is a clear indicator that persistent distortions in macroeconomic variables, such as interest and exchange rates, led the Government to forego a large source of foreign earnings. The general lesson for policymaking is that policies that combine both increases in domestic interest rates, as well as a reduction in exchange rate differences, can increase recorded remittances. Further, consistent government policies and an expansion in financial intermediation will help increase the size of recorded remittances and eliminate the parallel market.
Second Edition, Completely Revised and Expanded: Putting People First
Sociological Variables in Rural Development
Published for the World Bank by Oxford University Press
The second edition of this ground-breaking anthology, already widely used as a textbook, is to a considerable extent a new volume. Each contributor captures newly emerging trends in thinking and practice that are likely to characterize the 1990s. Some went back to the field to update cases and report new findings. Two new chapters have been added, and a more detailed index will help researchers and students alike.
Praise for the first edition of
Putting People First:
“An excellent book…perhaps the first of its kind. A definite contribution to the sociology of development.”—Sociological Bulletin
Please send me:
__ copies Paperback: 604 pages / Order Stock #60827 / US$19.95 / ISBN 0-19-520827-7
__ copies Hardcover: 604 pages / Order Stock #60873 / US$44.95 / ISBN 0-19-520873-0
Coupon orders accepted only from the U.S.A. and France. To order from other countries, request a list of World Bank Publications Distributors. Prices vary by country.
In the U.S. return the coupon to:
World Bank Publications
P.O. Box 7247-8619
Philadelphia, PA 19170-8619
In France return the coupon to:
World Bank Publications
66, avenue d’Iéna
75116 Paris, France