1. Bernard Weintraub, “U.S. Is Condemned on Salvadorans,” The New York Times, 21 May 1983, p. 5, col. 1. Quotas and a structured set of preferences apply to normal immigration to the United States. Special exceptions are provided under the Refugee Act of 1980, allowing immigrants to be admitted to the United States if they face persecution in their home countries or in the countries in which they habitually reside. The term refugee is applied to persons outside of the United States who convince U.S. officials abroad that they face persecution. Quotas set by Congress on the basis of administration recommendations govern the total number and the geographic distribution of refugee admissions. In practice, refugee status has been granted primarily to persons from Indochina. Central Americans almost never apply for or receive refugee status, their typical pattern being to enter the United States without inspection and apply for political asylum once here.

2. Ibid. See also Philip Shenan, “Salvadoran Refugees in New York Region Struggle for Asylum,” The New York Times, 25 July 1983, p. A1, col. 5.

3. “‘Young Male’ Salvadoran Documentation Material,” mimeo, National Immigration Project of the National Lawyer's Guild, Boston, Mass. See in particular the affidavit of Sandra Gutiérrez, 3 Nov. 1981, Redwood City, Calif. I interviewed staff members of organizations providing legal assistance to Central American refugees, selected from the Directory of Central American Organizations (Austin, Tex.: Central American Resource Center, 1984). I gave priority to groups in border areas because they have a better opportunity to observe the dynamics of the flow of Central Americans over time. Interviews were conducted by telephone, the number of interviews being constrained by cost. Interviewed were Juan Rascón of the Central American Refugee Project in Phoenix, 26 Nov. 1984; Jack Elder of Casa Oscar Romero in San Benito, Texas, 27 Nov. 1984; Linton Joaquín of the Central American Refugee Center in Los Angeles, 27 Nov. 1984; Claire Shurkowshi of Proyecto Libertad in Harlingen, Texas, 29 Nov. 1984; and Sister Rose Marie Cummins of Centro Presente in Cambridge, Mass., 9 Oct. 1984.

4. Americas Watch Committee, Guatemala: A Nation of Prisoners, (New York: Americas Watch Committee, 1984), 1–10.

5. Such an analysis would include examining patterns of political violence to establish whether poorer areas are more frequently attacked, whether the unemployed are more frequently targeted than the employed, whether economic conditions have deteriorated more rapidly in the worst areas of political violence, and which political and economic groups are most frequently targeted. Personal characteristics of victims of political violence could be compared with the Salvadoran population in general to identify groups that are particularly persecuted. Ideally, some sort of sample survey of Salvadorans outside of El Salvador would be conducted, although their lack of legal status in most countries would make this undertaking extremely difficult. The Moakley-DeConcini bill (H.R. 822/S. 377) under consideration at this writing calls for a combination of temporary legal status for Salvadorans and a study of the conditions from which they fled to come to the United States.

6. Patricia Weiss Fagen, Applying for Political Asylum in New York: Law, Policy, and Administrative Practice, New York Research Program in Inter-American Affairs Occasional Paper no. 41 (New York: New York University, 1984), 33.

7. Most economic indicators for El Salvador showed some decline from 1980 onward. GDP per capita in constant dollar terms has declined steadily, although the estimates of populations used to establish per capita performance do not acknowledge the large numbers of Salvadorans who are out of the country because the figures are based on projections from previous censuses and historical population growth rates. See United Nations Statistical Office, Monthly Bulletin of Statistics (New York: UN Statistical Office, 1985).

8. Weiss Fagen, Applying for Political Asylum, 33.

9. See Amnesty International, Report on Torture (New York: Farrar, Straus, Giroux, 1973, 1975), 220–21. See also the annual Amnesty International Report (London: Amnesty International), 1975–76, pp. 110–11; 1977, pp. 140–41; 1978, pp. 120–22; 1979, pp. 61–63.

10. See Socorro Jurídico del Arzobispado de San Salvador, Solidaridad: Boletín Internacional, no. 40 (15 May 1983):3–6; reproduced in Health and Human Rights in El Salvador (New York: Committee for Health Rights in El Salvador, 1983), 24–26.

11. Human-rights groups argue that the death squads have operated under the command of elements of the armed forces and in many cases have involved active armed forces personnel. This assertion has been vigorously denied by the U.S. State Department. But the dramatic drop in death-squad killings in response to U.S. pressure appears to confirm the argument that the death squads are under military command and control.

12. See Oficina Tutela Legal del Arzobispado, Comisión Arquidocesana de Justicia y Paz, Informe Anual 1982 (San Salvador), cited in Health and Human Rights, 27–29. See also Free Fire: A Report on Human Rights in El Salvador (New York: Americas Watch Committee and Lawyers Committee for International Human Rights, 1984), 3–5.

13. See Americas Watch Committee and American Civil Liberties Union, Report on Human Rights in El Salvador (New York: Random House-Vintage, 1982), 157–61. See also Free Fire, 1–46.

14. See Third Supplement to the Report on Human Rights in El Salvador (New York: Americas Watch Committee, 1983), 25.

15. Report on Human Rights in El Salvador, 196. See the testimony of Reps. Gerry Studds, Barbara Mikulski, and Robert Edgar, U.S. Congress, House Subcommittee on Foreign Operations, Hearings: Foreign Assistance and Related Programs Appropriations for 1982, 97th Congress, 1st Session, 25 Feb. 1981, p. 29. The Salvadoran army officers quoted are not named.

16. See Third Supplement to the Report on Human Rights in El Salvador, 117–18, 219. Eyewitnesses have identified the U.S.–trained Atlacatl and Ramón Belloso batallions as having perpetrated rape, torture, murder, and several massacres.

17. See Free Fire, 54–59.

18. Latin American Regional Reports: Mexico and Central America (hereafter cited as LARM), RM-81–02 (Feb. 1981), p. 4. Also Report on Human Rights in El Salvador, 158, 160; and Philip Wheaton, “Agrarian Reform in El Salvador: A Program of Rural Pacification,” in Revolution in Central America, edited by Stanford Central America Action Network (Boulder: Westview, 1982), 251–52.

19. El Salvador's Other Victims: The War on the Displaced (New York: Lawyer's Committee for International Human Rights and Americas Watch Committee, 1984), 29–31.

20. Weiss Fagen, Applying for Political Asylum, 38.

21. El Salvador's Other Victims, 31.

22. LARM, RM-81–05 (5 June 1981), p. 6. Report on Human Rights in El Salvador, 166–72. Third Supplement to the Report on Human Rights in El Salvador, 45–47.

23. National Immigration and Alien Rights Project of the American Civil Liberties Union, Salvadorans in the United States: The Case for Extended Voluntary Departure (Washington, D.C.: ACLU, 1983), 42–49. EARM, RM-84–02 (17 Feb. 1984), p. 5. Also, Linda S. Peterson, “Statement of the Representative of the Bureau of the Census,” in Central American Refugees, Hearing before the Subcommittee on Census and Population, Post Office and Civil Service Committee, U.S. House of Representatives, 27 June 1985, p. 7 (Washington, D.C.: U.S. Government Printing Office, 1985). Peterson estimates some seventy thousand Salvadoran refugees in Guatemala, although the male proportion is not reported.

24. See affidavit of Sandra Gutiérrez in National Lawyers Guild “Young Male” case materials, cited in n. 3.

25. See Shenan, New York Times article cited in n. 2.

26. Sergio Aguayo, “The Central American Exodus,” p. 9. This manuscript has recently been published as El éxodo centroamericano: consequencias de un conflicto (Mexico City: Consejo Nacional de Fomento Educativo, 1985).

27. Several of the staff members of refugee aid agencies heard reports from reliable Mexican sources that the U.S. Department of Justice has urged the Mexican government to take such measures in an effort to stem the flow of Salvadorans to the United States. If such pressure has been applied to the Mexican government, it contradicts repeated Reagan administration statements that Salvadoran refugees can remain safely in Mexico and therefore do not need to come to the United States.

28. “Monthly Report of Deportable Aliens Found in the U.S. by Nationality, Status at Entry” (October 1976 through June 1984), mimeo, Form G.23.18, available from Immigration and Naturalization Service, Statistics Division, 425 I St. N.W., Washington, D.C. 20536.

29. Weiss Fagen, Applying for Political Asylum, 33. Many INS apprehensions of Salvadorans take place not at the border but in raids of workplaces. The proportion of workplace apprehensions to border apprehensions varies considerably from month to month. Only border apprehensions (persons caught at entry or within seventy-two hours of entry) are used in this analysis. The percentage caught at the border is probably well below 25 percent, although the INS appears to be getting more effective.

30. Prior to 1982, a relatively small number of Salvadorans were able to enter the United States on nonimmigrant visas, which they regularly overstayed. Some of these individuals entered the United States by air. The vast majority of Salvadorans have entered without inspection via the Mexican border throughout the period examined by this analysis.

31. See Free Fire, 2.

32. See David Asman, “Behind the Human Rights Tallies,” The Wall Street Journal 10 Feb. 1984.

33. Figures for 1979 through 1981 were taken from the Socorro Jurídico tables reproduced in Health and Human Rights. Data were unavailable for three months—October through December of 1979. I used linear interpolation to fill in these missing observations, a reasonable procedure in this case because the level of death-squad violence showed only moderate variance during the first nine months of 1979. Furthermore, multiple sources cite a steady increase in death-squad killings during the final months of 1979 (the same trend obtained by interpolation), and the annual figure for 1979 from Socorro Jurídico matches closely that obtained using the interpolated figures. Figures for 1982 onward come from the five supplements to the Americas Watch and ACLU Report on Human Rights.

34. Supplement to the Report on Human Rights in El Salvador (New York: Americas Watch Committee, 1982), 12–13.

35. Ibid., 13. U.S. Reporting on Human Rights: Methodology at Odds with Knowledge (New York: Americas Watch Committee, 1982), 12–13; Free Fire, 1–46. Supplement to the Report on Human Rights, 106–7; and Supplement to the Report on Human Rights in El Salvador (New York: Americas Watch Committee, 1982).

36. Latin America Weekly Report (hereafter cited as LAWR), WR-80–07 (21 Mar. 1980), p. 12; LAWR, WR-80–13 (28 Mar. 1980), p. 1; LAWR, WR-80–32 (15 Aug. 1980), p. 8; Report on Human Rights, 157–60.

37. LAWR WR-80–42 (24 Oct. 1980), p. 1; LAWR, WR-80–44 (7 Nov. 1980), p. 2.

38. LAWR, WR-83–03 (21 Jan. 1983), p. 12; LAWR, WR-83–04 (28 Jan. 1983), p. 6; LAWR, WR-83–05 (4 Feb. 1983), pp. 6, 11.

39. LAWR, WR-84–01 (6 Jan. 1984), p. 4.

40. Another incident not represented by a dummy variable in the model was an operation of about five thousand men near the Chinchontepec volcano in the department of San Vicente during June 1980. Unlike other sweep operations designed to drive civilians out of FMLN-controlled areas (and thus deny the guerrillas food and other supplies), this sweep was conducted to ensure that no FMLN combatants remained in an already depopulated area being readied for the immediate settlement of thirty thousand persons previously displaced from other areas. See LAWR, WR-83–24 (24 June 1980), p. 4. I judged therefore that it was unlikely to have generated net out-migration in the same way as the other military sweeps represented in the model.

41. Free Fire, 54–57. According to State Department estimates, which may be biased upwards, ten to fifteen thousand persons have fled into Honduras to escape FMLN recruiting drives. How the causal relationship is established between this refugee flow and FMLN recruitment (as opposed to ongoing combat in and around “controlled zones”) is unclear.

42. Draining the Sea: Sixth Supplement to the Report on Human Rights in El Salvador (New York: Americas Watch Committee, 1985), 59–61.

43. See the interviews cited in n. 3. All staff interviewed had dealt with hundreds of Salvadoran clients, and all estimated travel time as taking from two weeks to three months.

44. Although the variance in the murder variable is high (the Pearson correlation coefficient between the level of murder and the same variable lagged one month is + .56), it contains sufficient multicollinearity to allow including only two, rather than all three, of these lags in a regression equation. If all three are included, lags 2 and 4 become insignificant, reflecting the high standard error of estimates caused by multicollinearity. The partial correlations of lags 3 and 4 with the dependent variable were the highest, so these two lags were used in the model presented. Unlike the dummy variables, the murder variable is not significant for the first two months (lags 0 and 1). This result holds regardless of whether lags 2, 3, or 4 are included in the equation, indicating that lags 0 and 1 do not belong in the model.

45. Free Fire, 2.

46. Initial model estimation was done using MicroTSP; the stepwise regression was performed with SPSS. The findings were verified using the TROLL econometrics system at MIT. The Durbin Watson statistic for the regression was 2.03, indicating very little probability of first order serial autocorrelation. I also checked for other kinds of autoregressive and moving-average processes and found none. Scatter plots of the residuals against each independent variable appeared homoscedastic. To address the seasonality of the Mexican and Salvadoran apprehensions series, I estimated the same model using seasonally adjusted transformations of these two variables. Adjustment was performed using the ratio-to-moving-average technique for both series from October 1976 through March 1984. The results of the regressions using the adjusted series are presented in table 3. They are consistent with the findings obtained without the seasonal correction.

47. Interview with Jack Elder, Casa Oscar Romero.

48. LARM, RM-84–05 (8 June 1984), p. 5. These figures came from Strategic Survey (London: Institute of International Strategic Studies, 1983.

49. See Indicadores sociales y económicas (San Salvador: Estadística y Censos y Ministerio de Planificación y Coordinación del Desarrollo Económico y Social) for the years 1977 through 1983.

50. Kirkpatrick, Jeane, “Dictatorships and Double Standards,” Commentary 63, no. 5 (Nov. 1979):34–45.

51. See the statement by Linda S. Peterson in Central American Refugees, p. 14