economic and quantitative analysis topics, roughly six in total. Those topics, in order, are focus on a non-core variable (and model) for the country of Nigeria, analysis of the World Bank World Development Indicators (WDI) model, comments on regression and the validity questions rising from within, the general problems and issues with regression analysis in general, whether any of the variable in the author's personal model have problems related to non-stationarity, and finally a suggestion of any changes to the models mentioned previously based on the outcomes that came about and the policy implications that can be garnered from this analysis.
Nigeria Non-Core Variable
The government and economics of Nigeria heavily focuses on exports of its vast mineral, petroleum and natural gas resources. An industry that has been a microcosm as compared to those exports over the recent years and decades has been the service sector (Nigeria, 2012). In many ways, Nigeria is the polar opposite of the modern-day United States, which has a higher (and growing) service economy while its manufacturing (especially of simpler goods) is falling rapidly.
The wider GDP of Nigeria (inclusive of both common staple exports like oil and minerals as well as everything else) has actually done quite well over the last 5-6 years ranging from 2003 to 2009 even with the global recession. The arc of exports has been steadily up (although 2006/2007 was flat and 2007/2008 was very flat). GDP took a drive from 2009 to 2009 but has also steadily rose despite this dip.
To expand on the model referenced in the prior section, it is noteworthy that Sub-Saharan Africa had a meager growth amount in 2008 but actually did not fall all that far in 2009, especially as compared to other areas of the world including eastern Asia, Europe, central Asia. This, combined with the fact that rising incomes cause poverty to go down (and vice versa), will cause the impact to the service sector as well as to poverty in general would be much more muted than it would be in other countries. The aforementioned pattern of rising exports buffering and helping the service sector in Nigeria (and the same in reverse) is supported by the WDI information even if the reactions would be much smaller than perhaps expected and thus it might be harder to attribute rises and falls (as small as they are) to export levels. However, some useful outcomes and analysis are certainly still possible.
The chart showing the relationship between GDP and exports, from 2003 to 2009, is shown in Appendix I. As noted before, both metrics trend sharply upward and largely in unison but the harmony between the two metrics is not completely constant. Exports from 2007 to 2008 was flat but there was a huge spike (more than $40 billion) in GDP. The exports rose the next year by about $16 billion but GDP actually fell nearly back to the 2007 level that same year. Even with this noticeable aberration, there is a clear correlation, if not clear causality, between these two variables when looking at the data from 2003 to 2009.
This data shows validity and reliability, at least to the extent that an external variable of import (like the global financial crisis) is not artificially influencing the numbers. It is expected by the author of this paper that minus such a seismic global economic phenomena, the odd figures from 2007-2009 would be mitigated if not entirely absent. As noted elsewhere in this report, there should generally be a correlation and causality between exports and GDP but tertiary metrics and events can skew this relationship. If one didn't know of the global financial crisis in 2007-2009, the odd numbers from those years would be quite puzzling, for sure.
As will be fleshed out much more in the next section, the problem with regression analysis is differentiating between causality and correlation. The first question that should be asked is whether it makes logical sense for gross domestic product to be affected by exports. The answer is, of course, a resounding "yes" and for two major reasons. First, exports are the bread and butter, so to speak, of the Nigerian economy. Because exports are the bulk of what Nigeria has to offer from an economic standpoint and if the core of its economy takes a blow, then all ancillary industries will be at least equally worse off as a result. It is similar to the argument, from an international perspective, that when the United States gets sick, the world gets a cold. When an economic heavyweight nation or sector stumbles, there can be wide-reaching effects and they can happen both domestically to a nation as well as internationally. When Nigeria oil exports are impacted, for example, it almost always has an effect on gas prices around the world.
As for the regression, it is clear that exports and GDP performance have a strong relationship, but it's not absolute. In other words, it is not a 1:1 dichotomy and there are other actors that can mitigate or aggravate a situation. A great example would be in countries like Mexico where the price of corn is impacting them greatly because of the high use of ethanol to fuel cars. To bring the discussion back to Nigeria, looking at things remotely close to a 1:1 perspective in terms of one variable affecting another is almost always simplifying it way too much and any validity of results is going to be affected by the fact that other events and situations can and do affect both exports and the GDP itself.
Also, the impacts between the two are mostly one-way. If a GDP stumble came about, the chances it would affect exports in any demonstrable way is very unlikely unless the GDP facet in question is something very vital and important to the export industry. Government policy and events, for example, can affect one or both of these two groups at the same time as well as individually.
As noted above, but will be fleshed out more aggressively in this section, the problem with aggression is that taking variables A, B (or even C. And beyond) is never going to be entirely definitive in most situations and make the research too complex will make any coherent and cohesive conclusions close to impossible. This morass is typically avoidable in quantitative-driven situations or anything else where an enclosed laboratory environment is possible and does not ruin the results. Drug research would be an example of something where the situations can be tightly controlled and a control and independent variable can be compared and contrasted.
However, economics is certainly a field where regression is exceedingly difficult because a laboratory environment is difficult to impossible to pull off and/or putting people in an artificial environment would not be remotely the same as having them out in the real world and oblivious to the fact that they are being assessed and measured. Looking at systems and structures eases this difficult and economics is certainly centered on that. However, those structures are ran by people and this thus leads to irrational and incorrect decisions being made. Greece can go on and on about how government assistance and the social safety net framework is important but it is quite clear that spending on those measures bankrupted that country or, at the very least, cutting those programs is the primary way by which the budget of that country will be brought back into check.
The analysis above about the exports and GDP performance in Nigeria certainly can and does have an effect on the performance and outcome of regression analysis but so long as ad hoc and other big-time variables are taken into account if and when they happen, the pitfalls of regression analysis as noted above can largely be avoided. For example, if there is a bombing and that affects exports and, by extension, GDP growth, that should be factored into the results. A bombing would actually reinforce the trend, in all likelihood.
Both exports and GDP would be non-stationarity in nature over time. The fact that exports are growing at a much faster clip than the GDP is, in and of itself, dictates that the distribution over time would necessarily have to change. Perhaps in a few decades, there will be a clear ratio between GDP and export levels, but until exports remain at a stable percentage of the Nigerian economy (and that is nowhere near happening yet), then this will simply not occur. The ratios that do exist (expected and actual) would have to change for any predicted results to be reliable and valid. Just looking at the data for this report shows a clear shift in ratio. In 2003, exports made up nearly 27% of the GDP of Nigeria. By 2009, that figure shifted to 44% and the immediately prior to that (2008), the ratio dipped back under 30% but immediately shot back up.…