Did border closure cause food inflation in Nigeria?
Ridwan Bello, August 2022
Ridwan Bello, August 2022
Welcome to the maiden episode of #CausalInfWithDataViz, a monthly series in which I use econometrics intuition, data science tools, exciting story-telling, and publicly available data to create visualizations and narratives that attempt to answer causal questions on issues of public interest. In this episode, the topic of discussion is the effect of a controversial trade policy on food prices in Nigeria.
FIRST, A BIT OF CONTEXT
In August 2019, the Nigerian government partially closed the country's land borders and, in October 2019, halted all trade via land borders. This policy, which remained in effect until December 2020, was aimed at stopping the smuggling of food items into Nigeria.
While the government bragged about the policy (e.g., see tweet below), many others decried it. One argument commonly peddled in public opposition to the policy is that the border closure led to food scarcity and rapid increases in food prices. But, is this true? Did the policy really cause food inflation in Nigeria? Let's dive into the data and find out!
BORDER CLOSURE AND FOOD SUPPLY
To begin, let's consider the nexus between the border closure and Nigeria's food imports. Figure 1 below presents the trend in food imports from 2011 to 2020 (the last decade for which data is available). The chart shows that, while food imports (as a share of total imports) had been dropping before borders were shut, it reached a 10-year low in 2019, the same year the policy took effect.
Note: Food imports comprise imports of foodstuffs, vegetable products and animal products. Data source: Observatory of Economic Complexity.
Next, let's explore Nigeria's food production data. Figure 2 presents the trend in the real (as opposed to nominal) value of Nigeria's aggregate agricultural output from 2011 to 2020. Here again, it looks like domestic food production has been on the rise since 2011 but reached its highest levels in 2019 and 2020, i.e., the two years in which the border closure was in effect.
Putting these two trends together, it looks like food import substitution, which had been ongoing in Nigeria pre-border closure, attained new heights post-border closure. Whether this is a genuine effect of the policy or a chance occurrence is hard to tell, but I will return to this point in a second.
BORDER CLOSURE AND FOOD INFLATION
Let's now consider the link between the border closure and food prices. To do this, we shall look at the Consumer Price Index data from the International Monetary Fund (IMF).
For the sake of readers unfamiliar with the concept, you may think of Consumer Price Index (CPI, for short) as a number used to measure (changes in) the price of a given basket of goods and services typically purchased by - guess who - consumers. CPI is measured relative to a base period, and changes in this index are sometimes used as a proxy for inflation. You can read more about CPI here.
I use this CPI data to visualize the trajectory of the price index for food and non-alcoholic beverages (as a collective basket of goods) in the period preceding, during and after the border closure. Causal inference diehards might call this an over-simplified Event Study, and they would be right. Those sanctimonious nerds!
Note: Figure 3 is fully interactive, and offers a slider that readers can use to adjust the view. But this figure may appear funny on mobile phones because of their small screen sizes. If possible, please view on PC or Tablet. Trust me, you will love what you see.
Two crucial observations stand out from Figure 3. One, food inflation increased instantly and consistently after the border closure took effect. To elaborate, food inflation hovered around 13% year-on-year pre-border closure (that means a given basket of food items cost about 13% more in any given month than it did in the same month of the previous year). But, after the borders were shut, food inflation immediately began to creep higher, surpassing 19% by December 2020.
Let me put that into context. Suppose that in July 2019, pre-border closure, you had successfully negotiated a 13% salary increase with your boss to help you cover rising food costs. By December 2020, that 13% bump in your salary would no longer suffice as you now needed a 19% raise on your original salary to cover your food purchases. Time to ask have another sit-down with your boss or find a better-paying job!
The second key observation from Figure 3 is that, following border reopening in December 2020, food inflation began to drop, but with a 3-month lag (Fun fact: this is what Economists mean when they say "prices are sticky downwards" - prices rise quickly but drop slowly). This second observation is another insight to which I return in a moment.
BRINGING IT ALTOGETHER
What do I make of all of these, you might ask? Well, Figure 3 clearly shows that the trajectory of food inflation changed following a change in the status of Nigeria's borders. And this happened, not once, but twice! If I were a lawyer and this was a legal case, that data, along with insights from Figures 1 and 2, would be a smoking gun. My opening argument in court would be that the border closure precipitated a reduction in food importation (as seen in Figure 1), but this reduced food supply from abroad was not sufficiently offset by the increase in domestic food production encouraged by the policy (as seen in Figure 2). The consequent reduction in aggregate food supply caused food prices to skyrocket (as seen in Figure 3). When the government reopened the borders, imported food began to flow in freely once again leading, within 3 months, to slower food price rises (again, as seen in Figure 3). Case closed!
But then, what do lawyers know about causality? 🤪 After all, they don't teach confounding or omitted variable bias or reverse causality or simultaneity bias in Law School. 🤪 (Hey, I'm kidding - in case you did not notice. Lawyers are brilliant, and I respect what they do.)
Seriously though, that hypothetical court case argument, while compelling, is not conclusively causal. This is because there may be other possible explanations - besides the aforementioned argument - for the trends that we have observed in the data.
Off the top of my head, I can think of a few of such alternative explanations. For example, it could be that commodity prices at the international market increased around the same time borders closed and reduced around the same time borders reopened such that the observed rise-and-fall trend in food inflation was a result of these commodity price changes rather than the border closure.
Fortunately, this commodity price explanation, when put to test, does not seem to hold too much water. Food commodity prices have generally been on an upward trajectory since early 2021 (see Figure 4) and rising food commodity prices at the international market could not have led to falling food inflation in Nigeria as observed in the months post-border reopening. So, we can probably rule out this explanation.
Yet another possible explanation is that foreign exchange fluctuations may, for some reason, have been correlated with border closure and may have been responsible for the observed patterns in food price index.
Look, it could even be that after the borders closed, all the food sellers in Nigeria decided to secretly come together, collude and take advantage of the situation to enrich themselves. This they did by artificially raising food prices after borders closed and lowering them after borders reopened!
That last explanation is a bit nonsensical, I'll admit. But, in my defense, when you are trying to prove causality in Econometrics, the standard practice is to consider as many unexpected but possible causal channels as you can. Sounds daunting? Well, welcome to the world of causal inference Econometrics!
WRAPPING UP
To conclude, I think that it is hard to prove, beyond all reasonable doubt, that the border closure caused food inflation. But, given the trends we've observed in the data, it is even harder to claim that it did not. If I were into deities, superstitions and that kind of stuff, at this point, I might consider consulting an oracle to be sure which way to swing on this one. Òrúnmìlà needs no Econometric evidence to diagnose causal effects. If, like me, you saw Nigerian movies growing up, you will understand. 😂