Evaluation of the methodological approach

This paper set out to implement an optimal approach to the regulation of Bitcoin.

Earlier research had indicated that the unique characteristics of Bitcoin meant that it served multiple use-cases at once and that the traditional approach to regulation, where an asset is first classified based on its singular use-case and then regulated by the singular appropriate regulator, was incompatible with this multi-use-case nature (Nabilou, 2019). Whilst this earlier paper indicated that applying regulation to multi-use-case assets on a use-by-use basis would provide the most complete regulatory approach, it stopped short of actual

implementation. In order to evaluate whether this approach could be feasible, an adapted method of regulatory effectiveness evaluation, as proposed by the OECD was implemented in this paper (Parker & Kirkpatrick, 2012). This approach involved the determination of the policy goals of the regulators, and the subsequent quantitative evaluation of the impact of different regulations on said policy goals to determine their effectiveness. This research paper therefore first determined the three primary use-cases of bitcoin and both the policy goals and regulations that the relevant regulatory bodies used for each use-case. Through regression analysis on the effects of these regulations on their relevant policy goals, it was possible to identify the regulations that had desirable effects on the policy goals and include these in the

sub-baskets of regulations. With the sub-baskets compiled, and the regulators’ policy goals considered, it was then necessary to evaluate how these regulations would interact with Bitcoin. In the introduction to this paper, it was determined that the optimal approach to the regulation of Bitcoin would see the formation of a basket of regulation that provides full achievement of regulatory policy goals, whilst maximising the preservation of Bitcoin’s unique advantages. The regulations from each of the sub-baskets were qualitatively evaluated and their potential interactions with each other, as well as how they influenced the advantages and disadvantages of Bitcoin for each use-case were considered. Of the sub-basket

regulations, one regulation was found to be unnecessary and was therefore dropped from the optimised basket. The remaining regulations were then combined into the final optimised basket of regulation, wherein each regulator implements the regulations that they contributed from their use-case.

It was discovered when reviewing existing literature that one of the greatest threats to successful regulation of non-traditional assets such as Bitcoin were competence limitations.

The threat of this was largely mitigated by the approach laid out in this paper, as, rather than designing new regulation for Bitcoin from the ground up, existing regulations were instead considered, enabling a more rapid implementation of regulatory policy that is already well understood by the implementing regulatory body. Furthermore, this deals with the additional issue that, whilst Bitcoin could be used as a reserve currency, a day-to-day currency, or an investment, it does not fit all of the characteristics of each of these use-cases. As a result, simply applying all of the existing regulation that applies to each use-case would not be desirable and indeed may not even be possible at all. Optimising based on regulators’ policy goals and then based on compatibility with Bitcoin ensures maximum regulatory

effectiveness with minimal regulation.

The final result of this paper, the successful creation of an optimised basket of regulation based on a practical application of the use-case approach to regulating Bitcoin proposed by Nabilou (2019), in conjunction with an adaptation of the OECD’s method of evaluating regulatory effectiveness, indicates that taking a decentralised approach to decentralised assets has merit for application to both Bitcoin and future multi-use-case assets (Parker &

Kirkpatrick, 2012).

Some issues were faced however, in the process of carrying out the research method. The first of these was discovered when attempting to gather data on the potential regulations for implementation; poor or inconsistent reporting standards. A lack of clear consistency between reporting standards meant that it was not possible to curate a specific list of regulations one by one, as the data available was not specific enough with regards to when exactly regulations were implemented. Furthermore, even when detailed descriptions of regulations were available, there was no framework that could be used in order to judge whether different countries’ regulations targeting the same area were similar enough to be considered as equivalent. Without this it was impossible to quantitatively evaluate the impact of implementing these regulations precisely. There is an irony in the fact that this paper, in attempting to overcome the lack of collaboration between regulatory bodies on a national scale, found that regulatory reporting on an international scale is similarly uncollaborative and disconnected.

To solve this, the research had to instead rely primarily on pre-existing regulatory

frameworks such as the Basel III banking regulation framework. This solution came with unique advantages and disadvantages of its own. Whilst the regulatory framework had accurate data on when regulations were implemented, as well as pre-vetted reporting

consistency, the number of countries for which data was available decreased dramatically as a result. Further to this, whilst the regulatory frameworks were designed to provide an exhaustive regulatory approach, there is still the potential for omitted variables. This was demonstrated to be true in the regression outputs as many of the regressions had low R-squared values in the region of 0.1-0.2. Whilst this is concerning initially, in real economic models it is rare to obtain high R-squared values for regressions. There are an almost infinite number of variables that could impact the policy goals being considered. The fact that this research included regulation from frameworks that were intended to be exhaustive, alongside additional control variables means that, whilst the R-squared of the models was generally low, the significant variables, and the inferences drawn from them are still valid. For future research to provide any deeper analysis, or to be able to include further regulations in the models, access to higher-level country specific data that is not publicly available would be required, for example, on the exact methods of implementing regulations and when those regulations were implemented.

There are some additional weaknesses to the approach taken in this study related to the assumptions made when performing analyses. Primarily, this is the reliance on the

assumption that all the regulations were implemented at different times. If this assumption didn’t hold, it may become impossible to distinguish between the different regulations. One method of minimising this risk was to include interaction variables to control for any

combined effects. This has additional importance, as some regulatory approaches may not be compatible with each other. When performing the results analysis however, in some cases attempting to include interaction variables led to the interaction term being omitted due to collinearity. No clear solution was able to be implemented in response to this issue, however, similar to other weaknesses mentioned above, more accurate data, for example providing exact dates when regulations were implemented would significantly minimise the risk that regulations are observed to be implemented at the same time.

The data limitations may have further, more significant implications for the final optimised basket of regulation. It is not clear whether the list of regulations included is exhaustive enough to cover all of the weaknesses of Bitcoin. A clear example of this is in Bitcoin’s significant energy consumption, much of which being produced by fossil fuels. Regression 6, evaluating sustainable development found a Market Abuse Directive would lead to more sustainable energy use, however, how exactly this occurs is unclear. It is possible, perhaps certain, that there are other regulations that would also have been found to significantly interact with the sustainable development policy goal or other policy goals under evaluation.

This then poses the question, is the optimal basket of regulation discovered by this paper truly fully ‘optimal’? Under the requirements of the approach, the intended method for optimisation was followed, balancing the policy goals of regulators against the weaknesses of Bitcoin. Optimisation was thus performed; it is simply more likely the case that the

regulations being optimised were not exhaustive. This may be because the data relied on panel data to build a generalised estimation of the effects of different regulations on the policy goals. Countries seeking to optimise their regulatory approach under a similar

methodology could use only data for their own country, providing more accurate results and higher quality data. I believe that with better quality and a greater quantity of data, more regulations would have been found to be significant that together would have truly covered all of the weaknesses of the Bitcoin medium. Therefore, this shortcoming is more a result of poor data rather than a poor model.