I think my first attempt at this was over-elaborate. I'm replacing it with this.

What's happening here is that they're trying to express what Alaniz et al. actually said in their article, which is as follows:

"We find that the minimum wage is high relative to the mean and median wages of private sector workers during the period that we study. The ratio of the mean minimum wage to the mean wage is 0.53 and the ratio to the median wage is 0.81."
http://ftp.iza.org/dp5702.pdf (p. 6)

What Alaniz et al. did was to collect data on real earnings in a sample of Nicaraguan households based on interviews conducted between 1998 and 2006, and compare these data with the minimum wages for the relevant sectors in each year (shown in their Table 1). From this they have calculated the ratio in their sample between the mean (average) real wage and the mean (average) minimum wage. The result is 0.53.

So then in "relación de 0.53 del salario mínimo con el salario promedio en promedio", the first "promedio" goes with "salario", meaning "mean wage", and "en promedio" goes with "salario mínimo", meaning "mean minimum wage".

"En promedio" really means "on average"; I think they mean to say that the ratio was 0.53 on average, though they don't make it clear precisely what they think "on average" means (the mean of the annual results?). All in all I think the safest option is to say what the source (Alaniz et al.) says.

--------------------------------------------------
Note added at 2 hrs (2015-09-21 06:51:13 GMT)
--------------------------------------------------

Or you could say "ratio of mean minimum wage to mean wage: 0.53".

--------------------------------------------------
Note added at 2 hrs (2015-09-21 06:54:25 GMT)
--------------------------------------------------

And as I said in my first answer, "para 1998 y 2006" is inaccurate. Alaniz et al's figure is not based on the two years 1998 and 2006 but on the period from 1998 to 2006, with data for each year in that period. I would be inclined to correct this discreetly, by putting "from 1998 to 2006" or "between 1998 and 2006", and add a note to point out the inaccuracy of the Spanish text.

--------------------------------------------------
Note added at 1 day14 hrs (2015-09-22 18:03:09 GMT)
--------------------------------------------------

Further to François' comment, there is no mention anywhere in Alaniz et al. of annual ratios. All the results they report are for the period as a whole, with no annual breakdown. The sole exception is Table A2, where they compare their dataset of real wages (annual figures from FIDEG) with the World Bank LSMS figures, based on a larger sample but only available for 1998 and 2005. (By the way, that must be the source of the confusion about the dates; "para 1998 y 2006" must be based on Alaniz's et al's reference to "the Nicaraguan LSMS survey carried out by the World Bank in 1998 and 2005", but they've garbled this, because the latter is used only as a comparison, not as the basis for the ratios cited.)

The overall ratios quoted are based on taking the ratio for each individual worker surveyed; they had to be, because the minimum wage in Nicaragua was different for each sector: "We assign to each worker in the FIDEG panel data set a minimum wage based on his/her industry of employment." In other words, it would have made no sense at all to take the annual mean of real wages reported and compare it with the annual mean minimum wage, without taking into account which was the relevant minimum wage for each worker surveyed.

--------------------------------------------------
Note added at 9 days (2015-09-30 09:54:26 GMT) Post-grading
--------------------------------------------------

I wasn't being serious :) And thank you for your patience and thoroughness!