Thank you, Mr. Chairman.
Thank you to the committee for the invitation to share with you the results of the research project that I've been involved with over this past year.
Basically, what we want to do through this project is to bring the numbers and to take them seriously. I have to say that all of the numbers I will be referring to are the numbers from the unions' NGO. In a sense, this bulletproofs the paper against any criticism of the source that we used to derive all of our results.
The debate about violence against union members, as you know, has been at the centre of a debate for the last few years.
Let me read a quote from a U.S. NGO:
Most of the violence against trade unionists is a result of the victims' normal union activities. While the Colombian government claims that most of the violence against trade unions is a byproduct of the armed conflict, the Escuela Nacional Sindical (ENS), a respected NGO that provides training and support to the Colombian labor movement, says that the majority of the anti-union violence that takes place in Colombia is in response to the victims’ normal union activities....
These are claims that we are going to test, using the evidence from the unions' NGO, the Escuela Nacional Sindical. Basically, in my view, the evidence has not been studied systematically to assess whether or not there has been progress in solving this issue.
There are some precise questions that we want to answer with this project.
First, what are the specific indicators of violence against union members?
The second question is a policy question: has there been any progress in solving this issue?
The third and fourth questions are going to be addressed using an empirical exercise. Can the killings of union members in Colombia be explained by their involvement in union activities, or is it really as a result of the armed conflict in Colombia that unfortunately they did not escape? As well, what are the main determinants of violence against union members in Colombia?
Basically, the research project is divided into two parts. The first part puts the figures in a few graphs and tries to answer the first two questions--namely, the evolution over time of the indicators of violence against union members and the different indicators of violence against union members, using always, or for most of the paper, data from the unions' NGO.
I will describe in a few words the empirical exercise. It's a little bit more technical and academic, but I'll tell you the main findings.
Independently of the data source used--from the unions' NGO, from the Central Unitaria de Trabajadores, which is a major workers federation in Colombia, or from the government--violence against union members in Colombia has steadily declined over the last seven or eight years. I can show you the figures.
Violence against union members has decreased at a steeper rate than violence against the general population. In fact, it has decreased 70% faster than violence against the total population. It has decreased faster than violence against other so-called vulnerable groups, which comprise teachers, NGO members, councilmen and former councilmen, and other groups.
The amount of government resources allocated to the protection of union members has increased steadily, and the number of union members protected has also increased steadily.
Finally, using different estimation strategies, time periods, and sources of information, we find no statistical evidence in support of the claim that violence against union members is caused by the involvement of union members in union activities. I will describe that in more detail.
Before I go through the stylized facts about violence against union members, let me tell you about the two figures that are most of the time brought up by people who oppose the free trade agreement.
The first piece they always bring up is that the number of killings of union members during the Uribe administration has been, I don't know, 1,700.
Let me give you an example. Suppose you have a country that has an inflation rate in 2000 of 30%, and by 2008 it has an inflation rate of 4%. You add up the inflation rates from 2000 to 2008 and you come up with a figure of 70%. You claim that this country has a hyperinflation problem, because over the last eight years it has had an inflation rate of 80%. You miss the trend if you take that number seriously. You should look at the yearly, monthly, or trimester indicators to gather information about the evolution over time of these figures. That's what I'm going to do when taking a look at violence against union members in Colombia.
The second piece of information given is that Colombia has 70% of the killings of union members in the world. The question that should be asked is who in the world reports the number of killings of union members. You start looking, and you see in the sample Sweden, Denmark, Finland, Canada, the U.S., Colombia, and other countries. But the only country in the world that is a developing country with a serious problem of violence that consistently reports to the ILO and other organizations the number of killings of union members is Colombia.
I was at a meeting one month ago of the ILO, and ILO staff recognized that Venezuelan workers and entrepreneurs were there. They reported that there were 152 killings of union members in Venezuela in 2008. If you put Venezuela into that sample of countries reporting the number of killings of union members, Colombia goes from having 70% of the killings of union members to having less than 12%. If you add Venezuela to the sample, Colombia decreases pretty quickly. If you're going to take the number seriously, every country should be reporting the number of killings of union members.
Let me go through the stylized facts now. You have the figures there. That is basically all data from the unions' NGO. There was a peak in 1996 in the number of union members killed. It decreased quickly until 1999 and increased again between 1999 and 2001. We reached another peak of 200 union members killed in 2001. Since then, the number of killings of union members has decreased steadily. The last year reported is 2008, when there were 48 killings of union members, according to the unions' NGO.
The natural question that follows is what the big thing of having that figure is if violence in Colombia has decreased in general. We divide the total number of union members killed by the total number of killings in Colombia, and we see that although both have declined, the number of union members killed has decreased 70% faster than the number of people in the total population killed.
The second piece of evidence, which is the standard way academic economists and criminologists measure violence, is the homicide rate. Basically, it is the number of homicides per 100,000 individuals. That's to take out the scale effect that is normally in these figures. In the stylized facts, we show that the homicide rate in Colombia for the total population was 70 in 2001, and last year it was 36. We construct the homicide rate for union members. That's the number of killings of union members per 100,000 union members. The homicide rate of union members in 2001 was almost 23, and by 2008 it was six. That homicide rate of six for union members is the same as the homicide rate in countries like Uruguay and the U.S. for 2008. So while the homicide rate in Colombia in 2008 was 36, for union members it was six. That's one-sixth of the total homicide rate in Colombia. Again, the homicide rate for union members decreases much faster than the homicide rate in Colombia.
Then, to my surprise, I have to say, if you compare the data from the unions' NGO, the Escuela Nacional Sindical, with the data reported by the government, the advances reached, using the unions' NGO data, are stronger than the advances reached using the government's data. In other words, the decrease in violence against union members is larger when you take the data from the unions' NGO than if you take the figures from the vice-president's office, which is the human rights office of authority in the government.
The human rights office of authority of the office of the vice-president also reports data on the number of killings of other vulnerable groups, which are journalists, NGO members, councilmen and former councilmen, and union members.
Basically what we see is that the number of homicides of union members has decreased faster than the number of homicides of these other so-called vulnerable groups.
The fifth piece of evidence is reported not by the unions' NGO or the government. It is directly reported by the Central Unitaria de Trabajadores, which is the largest Colombian confederation of workers. They basically report, case by case, the killings of union members, and they group the killings into union activists, union leaders, and unionized workers. So we basically take union activists and union leaders as a group of union leaders, in figure 5, or in stylized facts (V), and we show the evolution over time of the number of killings of union leaders in Colombia. In 2006 and 2007 there were no killings of union leaders reported by the CUT, but unfortunately, in 2008, there was a spike and they reported 11 killings of union leaders in Colombia.
The last piece of evidence is basically the amount of resources spent by the government in the protection of union members in Colombia and the number of union members protected in Colombia. Basically what we see is that in real 2009 Colombian pesos, the government was spending, in 2000, $5 per union member in protection schemes, and by 2008 the government was spending 100,000 pesos, which is about $50--$50 per union member in 2008 and $5 per union member in 2000. So there was a tenfold increase in the amount in real pesos. It takes care of the inflation rate and everything.
So basically there was a tenfold increase in the amount of resources spent by the government in the protection of union members during the last seven or eight years.
The natural question is, what is this money being used for? To protect the same number of union members, or is it that the government is also protecting more union members? That's answered by the last graph, which says that in 1999 and 2000 there were about 80 union members protected. This figure is hard to interpret because it's per 100,000 union members. Let me give you just the absolute figures.
In 1999-2000 there were between 50 and 80 union members protected, and last year there were 2,000 union members protected. That is 2,000 union members or leaders with protection schemes from the government, which consist basically of bulletproof cars in many cases—not in all cases—and bodyguards and security services for the person. And if it's a leader, he has protection with his family and all the things.
Those are the stylized facts about the evolution over time of the killings of union members in Colombia. Those are different indicators, different data sources, different comparable groups, etc.
Now for the second part of the paper, and I'm going to be brief on this. It's more empirical, more academic, and more technical, but let me tell you what we do.
We take a panel data set that contains data, per state and year, from 2000 to 2008, on the number of killings of union members, and we have data on union activities. We have a large variety of indicators of union activity, which we divide into two groups, or we put them together in some robustness checks.
Basically we call type 1 union activity formal types of union activity; that is, negotiations between the firms and workers about wage agreements and wage negotiations. We call that formal union activity. The informal or less formal union activity is what we call active acts of protest, which are street marches, strikes, hunger strikes, marches in the streets, etc., as active acts of protest by the unions.
In order to control for scale effects, we construct a measure of the intensity of union activity, which is the number, for instance, of strikes per union member in each state and each year from 2000 to 2008. So we have a relatively large data set in order to test the claim—and that's what we precisely want to do—that more union activity leads to more violence against union members. That's a statistical way of testing whether union activity is a dangerous activity in Colombia or not.
We also want to test what is causing violence against union members. We controlled for the level of economic development, that is the GDP per capita in each state and each year. We controlled for the general level of violence as captured by the total homicide rate for the general population. We controlled for a state presence, which is the number of police arrests per capita—it's a standard measure used in the literature to control for a state presence. We also controlled for the guerrilla and paramilitary presence—the number of attacks of guerillas and paramilitaries per capita.
Basically the claim we want to test is that the greater intensity of union activity has a causal impact on violence against union members. In the words of the U.S. NGO, “Most of the violence against trade unionists is a result of the victims' normal union activities”, or “...the majority of the anti-union violence that takes place in Colombia is in response to the victims' normal union activities....” If this claim is true, we should find a statistically positive effect of union activity on violence against union members.
The technical notes, all the details, all the data, and all the programs and codes we use are available for any verification anyone wants to do of the things we did.
The main findings are that we don't find any statistical evidence supporting the claim that violence against union members is caused by the activities of unions in Colombia.
What we do find, and I think this is crucial for the study of the free trade agreement, is that it is indeed true that in those states where there is less economic development and a lower level of GDP per capita, there is more violence against union members. In those states where there is more violence against the general population, there is also more violence against union members, unfortunately. In other words, union members do not escape the violence in Colombia, unfortunately. They are also victims of violence in Colombia.
Why do I say this is crucial for the discussion of the free trade agreement? All the details are in the paper, but using different estimation strategies, data sets, data sources, and time periods, we ran a large battery of robustness checks to be sure what we are saying is true throughout the exercises we've run. It is true that lower levels of economic development do indeed cause more violence against union members.
How can an FTA help? It can help basically by increasing the level of economic development in Colombia by bringing work and increasing the number of jobs in Colombia, increasing the economic development, increasing technological transfers through intermediate inputs that are traded in free trade agreements, etc.
This is the best thing, I think, that can be done to improve not only violence against union members, but in a discussion like this, people should also care about violence against everyone in Colombia and not just union members, not just union activists, and not just policemen. The best way to improve security in Colombia is by bringing opportunities for poor people to engage in work in formal working activities, bringing in more education, facilitating technological transfer, and promoting economic growth in general in Colombia.
To conclude, what we do in this paper is study the evolution and determinants of violence against union members. As I said, there are two main findings.
First, any indicators we used from the unions' NGO say very clearly that there has been a dramatic improvement in the security of union members in Colombia over the last seven or eight years.
Second, in this empirical exercise we didn't find any statistical evidence supporting the claim that in general violence against union members is caused by the involvement of the victims in normal union activities. I have to say very clearly that this does not neglect the fact that there might be cases of targeted violence against union members or NGO activists, etc. Although every rule has an exception, and exceptions are always brought up, I think it's also very handy to note the rule and understand the rule, and the rule in Colombia, according to recent statistical evidence, is that violence against union members is not caused by a greater intensity of union activity.
Thank you very much.