Imperial County is the fifth poorest county in the state. In 1998, Imperial County per-capita income of $17,353 was 62 percent of California's ($28,163) and 64 percent of the national average ($27,203). In spite of its relatively low per-capita income, Imperial County is actually better off than border cities in Texas. El Paso, Laredo, Brownsville, and McAllen, have per-capita incomes ranging from $15,750 to $12,330, and are the four poorest metropolitan areas in the nation.

Any discussion of the border, and of the successes and failures of the North American Free Trade Agreement (NAFTA), must take these facts into account. Indeed, NAFTA critics, and even some supporters, view the continuation of income gaps between the border region (outside San Diego) and the rest of the country as a sign that NAFTA has failed or is not working as it was supposed to. This issue of the Cross-Border Economic Bulletin focuses on the causes of low incomes in Imperial County. Its main findings are that:

• The income gap between Imperial County and California has been widening since at least 1969;

• Imperial County's large farm sector has declined in importance over the same period, and cannot explain the income gap;

• Most of the reasons for low per-capita income stem from labor market and human capital variables.

• Key variables are the level of educational attainment, the low rates of labor force participation, the county's youthful population, and the high rate of unemployment.

Income and jobs in Imperial County

In 1972, per-capita income in Imperial County was nearly 90 percent of the state level. This was an improvement over three years prior, in 1969, when it was only 81 percent. After 1972, however, the gap between the state and the county consistently widened, at the rate of about 0.8 percent per year (Figure 1).

The jobs picture in Imperial County over the same time is a mixture of stagnation and rapid growth. Job creation from 1970 to 1998 experienced relatively high growth from 1970 to approximately 1976, a decade of stagnation from 1976 to 1986, and higher than average growth since. Using the rate of growth of jobs statewide as a benchmark, California and Imperial County are compared in Figure 2.

During the first and third period, county employment growth was generally higher than growth at the state level. One can only speculate whether the pickup in job growth after about 1986 might have something to do with Mexico's economic policies, since it was in that year that Mexico joined the General Agreement on Tariffs and Trade (GATT) and shifted its trade and industrial strategies from inward-looking to outward-oriented.

The declining importance of agriculture

One of the primary differences in economic structure between Imperial County and the state is the much larger importance of agriculture. It is natural to wonder if incomes are low because of agricultural wages. This is an unlikely explanation, however, for two reasons. First, agriculture has declined in relative importance over the last quarter century. Measured in terms of county employment or income generated, its share of the county's economy has fallen over time. For example, in 1969, total employment in agriculture (farm production and agricultural services) was around 38 percent of the county's total, but by 1998 it had fallen to 26 percent. A second reason why agriculture cannot be responsible for the income gap is that it generates more income per worker than the average. In 1998, it generated just over 30 percent of county income, using 26 percent of the workforce.

Other sectors of the economy are equally unfruitful as explanations for lower incomes. Imperial County has fewer services-sector jobs and about the same share of retail-sector jobs, both of which pay below-average wages. On the other hand, the county has a very large share of its employment in government, which pays above-average wages. This more than compensates for a very small manufacturing sector where, again, wages tend to be above average.

Human capital and labor market variables

A better place to look for explanations is in the labor market, including the skills of workers and their opportunities to use their skills. In particular, four areas are critical. These are unemployment, labor force participation, the age of the population and education. Each one of these factors provide an important share of an overall explanation.

The unemployment rate in Imperial County averaged 29.6 percent between 1995 and 1999, and in October 2000, it was 27.5 percent, while at the state level it was 4.7 percent. It is beyond the reach of this short piece to look at how this number is determined, but it seems at odds with a number of other realities in the county, including the rapid job creation of the last 15 years. For example, during the depth of the Great Depression of the 1930s, unemployment in the United States reached 25 percent. This rate led to widespread farm foreclosures, defaults on home mortgages, no new investment, nearly zero new car sales, breadlines and so forth. None of this appears to be happening in Imperial County. Rather, it seems likely that the methods used by the United States Bureau of Labor Statistics to measure local area unemployment are biased in the case of rural counties with a large share of its labor force from across the border. In addition, the rate in Imperial County is much higher than in other California counties with similar economies. Nevertheless, even if the reported unemployment rate is double the actual rate, it is still very high. A reduction of the unemployment rate by 10 percentage points would add around 10 percent to total personal income in the county.

A second factor is the willingness and ability of the adult population to work. At the state level, the share of the adult population that chooses to work is nearly 7 percent greater than the share in Imperial County. In other words, not only does Imperial County have more unemployment in the group of people that want work, it also has a significantly larger share of its adult population choosing not to work. Rough estimates indicate that this reduces total income by about 6 percent.

A third factor is the share of the population too young to work. Imperial County is a young population and has about 5 percent more of its population in the 0-16 age group. All else equal, a population with more children will have lower per-capita incomes because each child is another person, but not another income. A rough estimate is that this reduces total income by about 6 percent.

Finally, and perhaps most importantly, incomes are lower as a result of the educational attainment of the county's population. Table 1 shows the figures from the 1990 census. While dated, they are not likely to improve in the 2000 census relative to the state or San Diego, given the very high rates of immigration experienced by the county. In addition, Imperial County has a very large share of its population that does not speak English. This limits the ability of residents to participate in the economy and reduces their incomes.

Recent analysis of the income effects of education show a wide gap between the wages of high school dropouts and those with more schooling. Table 2 illustrates these differences, using data from the 1996 Current Population Survey.

Putting these four elements together, a 10 percent reduction in unemployment (10 percent increase in income), 7 percent more of its adult population in the labor force (6 percent increase in income), 5 percent less of its population below 16 (6 percent increase in income) and educational attainment similar to the rest of the state (14 percent increase income), would increase total income in the county by about 36 percent. These estimates are admittedly crude, but they are an underestimate since they do not take into account the interaction effects of changes in the variables discussed. Nevertheless, as a simple exercise that begins to provide an explanation for the low incomes and poverty observed along the U.S.-Mexico border, this approach can explain a lot.