Dear Editor,
If critiques help to clarify ideas and concepts, if they cause us to reflect and change course, if they lead us towards the truth, that’s progress. This seems to apply to Joel Bhagwandin, even if he does not explicitly acknowledges it and instead conveys the impression he already knew his 8 October letter was based on a faulty premise. In his second missive, this indomitable analyst says that by the time his “response hits the media, part II was already published …” (SN, 11 October). That leaves the question unanswered: if JB knew migration cannot be estimated by arrival/departure data, what was the purpose of his first letter? That’s an apt question because JB even admitted the unsound assumption upon which his conclusion is based: “… in the absence of the net migration rate dataset, the net arrival/departure metric was used to extrapolate.” I read this statement as an admission by JB that he did not know his assumption was faulty and consequently could not have intended to write a follow-up missive. My critique opened his eyes.
One can only critique what one has seen and/or read, which is why my letter (SN, 9 October) was directed to JB’s first “essay about migration.’” If there is a second “essay,” I did not see it in SN, which is where the first one was published. I usually do not read the Guyana Chronicle and JB’s internet paper, “The Guyana Standard.” To set the context for the rest of the discussion, JB hypothesizes that migration (people leaving Guyana to live abroad permanently) is lower today compared to the “thousands” when the country was politically unstable and “bankrupt in the 1980s-90s” He used “airport” data to test the migration-hypothesis and assumed that arrivals and departures data are a good proxy for migration.
That assumption was incorrect and confusing as well: JB used the word “arrival(s)” 7 times, “departure(s) 5 times and “migrate/migrating” 6 times in his 8 October letter. One is left to wonder which is which. He concluded: “More importantly, this analysis of the immigration data for the period 2015-2023, has empirically disproven the implicit notion that a large number of Guyanese are migrating.” The next sentence after this quote was more specific and clearly demonstrates that JB uses the difference between arrivals and departures to conclude “that … thousands of persons are coming to and/or returning to Guyana annually, nearly 10,000 annually, more than those who are perhaps leaving the country permanently.” The only way to interpret the last four words of this quote is that JB used arrival/departure data to gauge migration. That’s a statistical non sequitur. In his second letter, JB speaks about “part I of my essay about ‘migration’” and observes that he never said net arrival/departure “is equal to net migration.” It is irrelevant whether or not JB was referring to migration or net migration; he simply cannot use net arrival and departure data to talk about migration, gross or net.
To dodge my critique, JB resorted to a non-sensical argument response, “extrapolation” (SN, 11 October), which, according to him, is a technique “employed when there is limited and/or the absence of precise and more granular dataset (s).” Extrapolation is NOT a research “technique.” It is a method to arrive at an educated guess of the value of a variable at a future time, say 2024, 2025 or any other year for that matter. The fundamental point is that while one can extrapolate the value(s) of a given variable, say arrivals in Guyana, at some future date(s), it is incorrect to equate that variable and its extrapolated value(s) to another variable, which in this case, is net migration. That’s not extrapolation or inference; it’s ignorance. In fact, there is no need for him to extrapolate airport arrivals and departures because his sample period covers 2015 to 2023 and he has data points for each and every year. Extrapolation is not a “weave,” jumping from one thing to a completely new thing. Extrapolation is simply a way to figure out the value of a variable at a future date.
Now let’s address issues he raised about my method to estimate migration in “Essays,” which was published in 2022. Here’s what JB wrote: “… Dr. Gampat asserted that the net arrivals dataset was inappropriate to derive the conclusion I arrived at … Dr. Gampat may have contradicted one of his own, old academic work on the very subject of migration. In his self-published book… he utilized the same metric to make the same conclusion, but he did so incorrectly, in that, he conflated the two concepts.” Either JB misunderstood or misrepresented – or perhaps both – my methodology and conclusion. Permit me, dear Editor, to clarify.
The Chapter on “The Guyanese Diaspora” attempted to “estimate the size and ethnic composition of the migrant Guyanese stock” (p. 516). Since Guyana does not publish data on migration (in and out), I used the Beautiful Population Identity, discussed at the opening of the Chapter, to estimate net migration. Because of the fundamental importance of this identity, let’s restate it: population change = natural increase plus net migration (PC = NI + NM). Since data are available for two of the three variables (population change and natural increase), I proceeded to test whether the difference between arrival and departures (airport data) would satisfy the population identity by acting as a substitute for net migration. “It [the testing] begins by using the BoS method of estimating net migration and then moves on to test whether the population identity is upheld or violated” (p. 527).
While the test was done for several periods, focus on the 19 years from 2000-2018 because that’s the period JB mentioned in his letter. Based on the arrival and departures data (see “Essays” for sources), net migration was 117,051 and natural population increase was 193,279. These two data point mean that PC = 310,330 during these 19 years. That was not the case: the country’s population fell by 25,262. Since PC and NI are official data, then NM is incorrect, which means that the difference between arrival and departure is not NM. That’s the purpose of Table 19.3 (p. 531-2): to demonstrate that when the difference between arrival and departure is plugged into the identity, it violates the integrity of the “Beautiful Population Identity,” and this true for all years and periods from 1960 to 2018.
The only way to restore the integrity of the identity is to accept the official data for PC and NI and solve for NM. With PC = 25,662 and NI = 193,279, then NM = negative 218,901 from 2000 to 2018. If the entire 59 years from 1959 to 2018 is considered, PC = 185,892 and NI =867,231. Solving for NM, the results is negative 681,429. These negative amounts represent the size of out-migration.
Let’s summarize: (i) it is incorrect to use the difference between arrival and departure data as equivalent to net migration; (ii) our critique of JB’s original letter (SN, October 8) made him changed tactics without explicitly acknowledging the accuracy of the critique; (iii) defending his position, JB says he extrapolated from arrival and departure data to gauge net migration which is nonsensical; (iv) if these were the case, there was no need for a second “essay;” and (iv) JB distorted how I arrived at net migration in “Essays.” The fundamental point is that I used the difference between arrival and departure data to demonstrate that it cannot serve as an estimate/substitute of net migration because it violates the Beautiful Population Identity.
Sincerely,
Ramesh Gampat