Article Text
PDF
Abstract
This paper critically analyses contrasting estimates of Malaysia’s illicit cigarette trade in 2011, 2015 and 2019 by Bui et al and Koya et al who previously produced independent estimates at about the same time using tax gap analysis. Collaboration between the two authors’ teams emerged due to the discrepancies in their results, generating this paper to explore the methodological issues identified and hence produce revised estimates of the rate of illicit. Key issues identified were: Bui et al’s assessment of legally imported cigarettes impacting all years; their exclusion of ad valorem duty affecting the 2011 and 2015 estimates; Koya et al overlooked the value of cigarettes for export market in their ad valorem calculation and used the sales value of imported tobacco/tobacco products, not just cigarettes, both of which impact estimates for 2011 and 2015. Recalculations using Koya et al’s consumption data reveal that in 2019, illicit cigarettes accounted for about 70% of the market, which is higher than Bui et al’s estimate (38%) but slightly lower than Koya et al’s (72%). For 2011 and 2015 where ad valorem applied, the corrected estimates show a share of the illicit cigarette market of approximately 41.1% and 52.7%, respectively, differing from Bui et al’s 0% in 2011 and 29.6% in 2015, and Koya et al’s 51% in 2011 and 55% in 2015. This paper provides essential lessons for addressing methodological issues between authors’ teams and updated estimates of Malaysia’s illicit cigarette trade, verifying that Malaysia faces a substantial illicit cigarette trade problem.
- economics
- illegal tobacco products
- low/middle income country
- taxation
- tobacco industry
This is an open access article distributed in accordance with the Creative Commons Attribution Non Commercial (CC BY-NC 4.0) license, which permits others to distribute, remix, adapt, build upon this work non-commercially, and license their derivative works on different terms, provided the original work is properly cited, appropriate credit is given, any changes made indicated, and the use is non-commercial. See: http://creativecommons.org/licenses/by-nc/4.0/.
Statistics from Altmetric.com
What this paper adds
Identification of significant methodological issues in two recent studies of the illicit cigarette trade in Malaysia.
Updated estimates of the illicit cigarette trade in Malaysia which verify the large scale of the problem.
The importance of collaboration between groups of authors to identify and correct methodological discrepancies.
Introduction
In their study entitled Magnitude of illicit cigarette trade in Malaysia: empirical evidence compared with industry studies, Bui et al provided estimates of the illicit cigarette market in Malaysia.1 Using a tax gap analysis methodology, they compared self-reported consumption (ie, as reported in national surveys) and tax data related to the use of cigarettes in Malaysia for the years 2011, 2015 and 2019, with the gap between the two data points indicating the extent of the country’s illicit trade. These estimates were then compared with estimates of the illicit cigarette trade provided by the tobacco industry. Bui et al’s study revealed a concerning level of illicit cigarette trade in Malaysia in recent years. The proportion of all cigarettes consumed that were illicit (based on their central estimates) was determined to be 0% in 2011, 29.6% in 2015 and 38.2% in 2019. It is noteworthy that these figures were lower than the estimates provided by the industry.2–4
While Bui et al’s paper was in development, Koya et al had also engaged in a separate study to produce estimates of Malaysia’s illicit cigarette trade, with both papers being published at about the same time.5 Koya et al’s paper used a similar methodological approach to Bui et al, using a tax gap approach and many of the same sources of official data. Koya et al’s estimates, however, were considerably larger and more in line with those put forward by the industry—51% in 2011, 55% in 2015 and 72% in 2019. Upon discovering the discrepancy, Koya et al began exploring the methodologies of both papers to identify the source of the divergence. It was quickly apparent that some small differences were inevitable given the minor variations between the two studies (eg, assumptions around the consumption estimates adopted or the number of decimal points used with the data). However, the nature of the larger disparities between the results of the two studies required a more substantive explanation. Having identified some possible issues, Koya et al reached out to the author team of Bui et al to gain the best possible understanding of the methodological approach adopted therein. This contact and the resulting discussions ultimately led to the creation of this paper.
Through a collaborative process between the two sets of authors, two fundamental issues were identified in the methodology employed by Bui et al that contributed to their lower illicit cigarette trade estimates. These concern how the size of the legal sales of imported cigarettes was assessed, and the exclusion of ad valorem duty when assessing the size of duty paid (ie legal) sales. The discussions also identified a methodological issue within Koya et al, in the way that the ad valorem duty calculations were conducted within that paper, and also the possibility of making an improvement in the application of Koya et al’s methodology since Bui et al had access to more accurate data on the value of imported cigarettes. This paper will explore all of these issues and, in doing so, demonstrates the value of collaboration in addressing methodological differences between two groups of authors. As such, the paper also presents updated estimates of Malaysia’s illicit cigarette trade, which corrects the methodological issues identified.
To begin, in the first section, we provide some relevant background information and an overview of Malaysia’s cigarette excise tax regimes from 2011 to 2019. In the second section, we take a step-by-step approach to explore the aforementioned methodological issues and ultimately update the tax gap calculations. Our initial step involves exploring and addressing the issue of how imported cigarettes were assessed in Bui et al’s study, using data from 2019 since only this methodological issue is relevant to the 2019 calculations. In the following section, we then explore the years 2011 and 2015, where we not only correct for the assessment of imported cigarettes but also address the issues surrounding the need to use ad valorem duty when calculating the tax gap. We then combine these results to present our revised estimates for Malaysia’s illicit cigarette trade and compare these with the estimates originally presented in Bui et al and Koya et al which will therefore summarise the impact of the tax gap methodological issues identified. Finally, we conclude by reflecting on the process of collaboration between author teams.
Background
The rate of illicit cigarette trade, both here and in the two original studies, is estimated through an assessment of the cigarette tax gap. The tax gap is quantified by comparing the tax revenue actually generated with estimates of the tax that should have been paid if all cigarettes were purchased and taxed legally (usually measured by multiplying the tax rate by estimated total cigarette consumption). Hence, to be able to calculate the tax gap, we must first understand the structure of the cigarette excise tax. The cigarette excise tax regimes implemented in Malaysia during the specified period of study consisted of two types—a specific duty system (2015–2019) and a mixed system consisting of specific and ad valorem duty (2011–2015).6
Between 2011 and 2015, the ad valorem component of the tax was 20% of the value of imported cigarettes (measured as the cost, insurance and freight (CIF) value for imported products)7 and/or the value of locally manufactured cigarettes (measured as the ex-factory value for locally manufactured products) minus export values (measured as free on board (FOB) value for exported products).8 It is important to note that the export values need to be subtracted from the calculation as not all imported and locally manufactured cigarettes were intended for the Malaysian market. In 2011, cigarettes were also subject to a specific excise tax rate of 0.22 Malaysian ringgit (MYR) per stick, along with the ad valorem duty. Between January and October 2015, the excise tax rate applicable was 0.28 MYR per stick together with the ad valorem duty. However, from November 2015, the duty changed to being a fully specific tax rate of 0.40 MYR (which remains the case at the time of writing).
For cigarette consumption, both Koya et al and Bui et al based their estimated figures on the official National Health and Morbidity Survey, and their findings are generally aligned.9 However, in this paper, we will exclusively use Koya et al’s cigarette consumption estimates for the years 2011, 2015 and 2019 due to some concerns surrounding the accuracy of the 2015 estimates derived from the survey data, as this year of the survey was found to be inconsistent and unlikely to be reliable (which was confirmed to be the case by relevant government ministries).5 Notably, Koya et al have accounted for these issues, whereas Bui et al did not. The study by Koya et al provided various estimate scenarios to accommodate potential inaccuracies in consumption surveys (as did Bui et al) but for the purpose of this study, only the central estimates from Koya et al are used for simplicity.
2019: discrepancy in assessing legally imported cigarettes by Bui et al
The first discrepancy issue when calculating the tax gap relates to Bui et al considering all legally imported cigarettes to be part of the legal retail market. For cigarette imports, Bui et al relied on data from the Department of Statistics Malaysia (DOSM), which provided data on the ‘volume (in kg) of legally imported cigarettes’ for domestic sale. In contrast, Koya et al calculated the volume of legally imported cigarettes by dividing the actual tax revenue from cigarette imports by the applicable excise tax rate. Since excise duty is only paid when imports are released for sale into the domestic market, any leakage of legal imports into the illicit market before official release for sale creates a disparity between the actual quantity imported and the duty paid quantity actually released for sale.10 Therefore, by calculating imported sales based on the duty paid, Koya et al did not assume that all legally imported cigarettes end up in the legal retail market. Furthermore, since both studies used excise duty receipts to measure the sales of locally manufactured cigarettes, adopting the same approach to consider both locally manufactured and imported cigarettes represents a consistent approach to measuring total legal sales volumes.
To illustrate the importance of using excise duty receipts for assessing the sale of imported cigarettes, we sought to develop estimates that correct for the methodological error. The first step in this process is to divide the excise duty paid on both imported and locally produced cigarettes by the applicable specific duty rate, to identify the relevant number of legal cigarette sales (table 1).
Calculations of the volume of cigarettes (number of sticks) for both locally manufactured and imported into the country in 2019
Using the duty paid approach, imports were identified as being 6.30 billion sticks. This is in stark contrast to the case of Bui et al where imports were reported as being 13.55 billion sticks based on the volume of imports, which therefore implied a far higher total level of legal sales than is in fact the case. This considerable difference suggests that there may be issues in controlling the tobacco supply chain in Malaysia, but exploration of such issues falls outside the scope of this paper.
Using the revised number of legal stick sales, we can compare the total number of cigarette sales with consumption estimates adopted from Koya et al’s study to compute the proportion of illicit cigarettes (table 2).
2019 tax gap calculation
By using the tax paid to estimate the size of legal imported sales, we find that illicit cigarettes occupied approximately 70% of the total cigarette market in 2019. This proportion is almost double that reported in Bui et al (38%), although they did use marginally higher consumption estimates. The tax gap presented in Koya et al (72%) is more in line with these correct estimates. One explanation for the difference is that Koya et al used revised government estimates of tax duty revenue instead of the actual data used here, as the revised estimates were the best available during the development of their paper.5 Given the above results, in 2019, most of the differences between the findings of Bui et al and Koya et al can therefore be attributed to their approach to measuring the sale of imported cigarettes.
2011 and 2015: discrepancies in assessing the ad valorem data and associated calculation methods
The second discrepancy issue revolves around the treatment of ad valorem duty in both Bui et al and Koya et al while estimating the illicit cigarette trade in 2011 and 2015. When calculating the total amount of legal cigarette sales in the domestic market in 2011 and 2015 (by dividing the tax revenue with the applicable excise tax rate), Bui et al only considered the specific element of excise duty, and therefore did not account for the ad valorem duty element that existed for the entire year of 2011, and from January 2015 to October 2015. The impact of not accounting for the ad valorem element of the tax is that Bui et al overestimated the number of cigarettes sold in the legal market and hence underestimated the extent of the illicit cigarette trade. This methodological issue was not present in Koya et al’s study. However, upon comparing the methodologies employed in both studies concerning the link between duty paid and sales, it became evident that Koya et al had overlooked the value of cigarettes for the export market. As previously explained, not all sales are intended for the Malaysian market; hence, export values should be excluded when calculating the ad valorem duty paid. This omission would cause an overestimation of ad valorem duty, and hence an underestimation of the legal cigarettes actually sold. This issue would be of particular concern in 2011 because it was before Philip Morris International closed its tobacco factory in Malaysia in 2012, followed by British American Tobacco and Japan Tobacco International in 2015.11–13 Furthermore, since Koya et al’s study used publicly available official government data, they had previously used the sales value of imported tobacco/tobacco products instead of only the sales value of cigarettes (which is not publicly available) to calculate the ad valorem element of the duty paid on imports. Using the larger value for tobacco imports would have also overestimated the amount of ad valorem duty paid on cigarette imports and hence created a further underestimation of the number of legal cigarettes sold. Such underestimation of legal sales would create an overestimate of the extent of the illicit cigarette trade.
Considering the methodological and data-related complications mentioned above, we recalculated the estimates of the illicit cigarette trade for 2011 and 2015. However, the presence of the ad valorem duty imposed in both years adds complexity to the calculation of the tax theoretically owed, and hence the method used above for 2019 needs a further step to account for the ad valorem element of the tax when calculating the tax gap for the years 2011 and 2015. Furthermore, the change in the system of duty during 2015 creates another challenge due to the adjustment in the tax structure for the last 2 months of the year.
Methodology
In order to explore the impact of the methodology issues described above and hence calculate revised estimates, we first need to ascertain the key data points that are required for the ad valorem element calculation. These key data points are the trade data, which include the value of imported cigarettes, the value of locally manufactured cigarettes and the value of exports. Since data on the value of imported cigarettes (represented by the CIF value in accordance with the standardised HS Code 240220900) are not publicly available, the data for the years 2010–2016 were sourced from the DOSM Trade Statistics by Bui et al who purchased it. The value of locally manufactured cigarettes (measured by the ex-factory value) from 2010 to 2016 is publicly available, hence was sourced from the DOSM Manufacturing Report. The value of exports (represented by the FOB value in alignment with the standardised HS Code 240220900) for the years 2010–2016, which is also publicly available, was extracted from the DOSM Trade Statistics.14 In order to allow for year-to-year changes in inventory stock movements and delay/mismatch in reporting, we calculated a 3-year moving average on these trade data. A 3-year average was chosen in light of the industry-reported 2-year shelf-life of tobacco products.15
We calculated the ad valorem component of duty using the method described in Koya et al. The challenge with ad valorem taxation is that the consumption estimates used are lacking in the sales value needed to determine the implied ad valorem duty. To address this, we initially subtracted the proportion attributed to the ad valorem component of the tax, from the total excise duty paid. We do this by calculating the total value of sales (by combining both value of imported cigarettes (CIF) and value of locally manufactured cigarettes (ex-factory), then subtracted the export value (FOB)) before applying the 20% rate to derive the estimated ad valorem revenue (table 3).
2011 and 2015 CIF values for imported cigarettes, ex-factory values for locally manufactured cigarettes, FOB values for exported cigarettes and ad valorem values
The estimated ad valorem duty was subsequently subtracted from the total excise duty paid, thereby revealing the remaining amount as the specific duty paid. The specific duty paid was then calculated as a proportion of the total excise duty actually paid. Next, to calculate the tax gap, we had to first calculate the specific duty theoretically owed by multiplying the estimated consumption with the specific tax rate. This amount was then uplifted by the ratio of ad valorem to specific duty identified in actual payments (as previously stated) to estimate the overall duty theoretically owed (since the hypothetical value of such sales is not known and would be needed to calculate the ad valorem duty to be paid). This overall duty theoretically owed was then compared with the total excise duty paid to produce the tax gap. To enhance the clarity of these methods, we have included concise formulas within the column titles for the various calculations in our comprehensive results in table 4.
2011 and 2015 updated tax gap calculation
Results
Applying the method outlined above to assess the tax gap (table 4), the 2011 estimated share of the illicit cigarette market is 40.7%, considerably higher than Bui et al’s estimate (0%, where they failed to accurately assess legally imported cigarettes and did not incorporate the ad valorem element in their calculation) and lower than Koya et al’s central estimates (51%, where they failed to deduct export values when calculating ad valorem duty and had to use the value of imported tobacco instead of only cigarettes). When the recalculated tax gap incorporates the use of a 3-year moving average, it sees a slight increase to 41.1%.
For the year 2015, we present the tax gap in two scenarios. Scenario I assumes that a single duty rate applies for the entire 12 months, specifically the lower duty rate from the first 10 months (0.28 MYR). Scenario II distributes annual consumption evenly across 12 months, and then uses the lower duty rate of 0.28 MYR and 20% ad valorem for the first 10 months, and the higher duty rate of 0.40 MYR for the remaining 2 months. In scenario I, the updated estimates show that the share of the illicit cigarette market is substantially higher at 52.5%, compared with the 29.6% reported by Bui et al but are slightly lower than Koya et al’s central estimates of 55%. The estimates in scenario II show a higher share of the illicit market at 55.0%, aligning with Koya et al’s findings. When the 3-year ad valorem calculation is applied, the 3-year moving average indicates that the share of the illicit cigarette market is little changed, at 52.7% for scenario I and 55.5% for scenario II.
Based on the results presented above, it can be seen that the impact of Koya et al using the sales value of imported tobacco and not accounting for export value of cigarettes when considering ad valorem duty is less important in 2015 compared with 2011. This is likely due to the closure of three tobacco factories in Malaysia by 2015, thereby limiting their export opportunities, and the need to import tobacco leaf (hence the size difference between the value of imported tobacco overall and the value of just cigarettes). Conversely, the discrepancies in assessing legally imported cigarettes and the omission of ad valorem duty in Bui et al impact the tax gap calculations for both 2011 and 2015. The ad valorem issue and its components’ omissions, however, do not impact the 2019 estimate given that there was only specific duty applicable.
Overall comparison of tax gap estimates
Drawing together these revised estimates for 2011, 2015 and 2019, we can see how correcting for the aforementioned methodological issues in both studies has changed the overall picture of the tax gap, hence the extent of the illicit cigarette trade in Malaysia, as summarised in table 5.
Summary of corrected tax gap estimates for 2011, 2015 and 2019
Conclusion
In summary, this paper has shed light on the contrasting estimations of Malaysia’s illicit cigarette trade, as presented by Bui et al and Koya et al. It establishes that the latter’s projections appear to offer a more accurate representation of the trade’s extent, although with a tendency to overstate, particularly in 2011. The revised estimates here correct for the identified errors and hence are more reliable than either of the previous two studies. Nevertheless, it should be recognised that both Bui et al and Koya et al provide value by confirming that Malaysia has a substantial illicit cigarette trade problem, and with this paper providing important context for the estimates featured in them, they remain useful additions to the literature base.
The discussions that took place between the author teams that ultimately resulted in this article exemplify the power of collaboration in deepening our understanding of complex subjects. Our experiences here show that collaborating directly with a wider team of experts can sometimes be a very effective way of identifying any methodological and/or data issues that are unlikely to be picked up during the peer review process.
Ethics statements
Patient consent for publication
Ethics approval
Not applicable.
Footnotes
Twitter @JRBranston, @AllenGallagher_, @WencyBui
Contributors Paper formulated by all authors. RKK prepared the first draft with JRB and AG, and then WKTB, HR and NMN provided critical feedback, contributing to improvement of the final paper. The final version agreed on by all authors. The responsibility for any errors remains entirely with the authors. The corresponding author, JRB, stands as the guarantor, accepting full responsibility for the work and/or the conduct of the study, had access to the data and controlled the decision to publish.
Funding JRB and AG receive funding from Bloomberg Philanthropies as part of the Bloomberg Initiative to Reduce Tobacco use. RKK is supported by a University of Bath research Studentship and by a Central Bank of Malaysia PhD Scholarship. WKTB, HR and NMN have no funding to declare for this work.
Disclaimer The funders had no role in study design, data collection and analysis, decision to publish or preparation of the manuscript.
Competing interests JRB owns 10 shares in Imperial Brands for research purposes. The shares were a gift from a public health campaigner and are not held for financial gain or benefit. All dividends received are donated to health-related charities, and proceeds from any future share sale or takeover will be similarly donated.
Provenance and peer review Not commissioned; externally peer reviewed.