An Economic Perspective on Terminal Dues

Abstract

The purpose of the UPU terminal dues system is to compensate the designated operator in the destination country for the cost associated with the handling, transport and delivery of cross-border letter post items. However, the amount of compensation for each bilateral mail flow is not a direct translation of the costs, nor the price of equivalent domestic services. This results in several market distortions in the international mail market and financial transfers between designated operators. This paper investigates the impact of a proposed update of the terminal dues rates on the financial transfers between postal operators.

Appendix 1—Supplementary Description of Methodology

1.1 Description of Data Collection and Calculations

Bilateral mail flows

Bilateral mail flows between designated postal operators are an essential input in the model. In combination with the difference between the actual terminal dues rates and the counterfactual ones (i.e. the equivalent domestic postage), the magnitude of bilateral mail flows will determine the size of the financial transfers. Estimation of financial net transfers in our model requires information about bilateral flows of letter mail up to 2 kg between all designated postal operators (both in number of items and in weight), split by letter format (P, G, E). This information is not readily available in the public domain, but the data can be constructed our own data set.

Whereas we have information about flows of cross-border letter post between regions, we do not know how the cross-border volume to/from each region is disaggregated among the countries in the region. Moreover, we do not have information about the product mix (i.e., letter formats) in the volumes. For this reason, we use a number of proxies to approximate mail flows subject to UPU terminal dues.

As a first step, we create estimates of the total inbound and outbound mail flow (in terms of number of items) for individual countries. The starting point for this is readily available data for total inbound and outbound mail flows from the UPU (see UPU 2015) measured in number of items. Since 2011 is the year in the UPU statistics for which we have the broadest coverage of data (139 countries and territories covered), we use this as a reference. If we cannot find information for a specific country or territory in the UPU dataset from 2011, we turn to UPU statistics from other years. This adds data points for another 31 countries. When neither of these sources can provide us with an estimate, we turn to domestic sources such as the national regulatory authorities. This adds data points for another five countries. For seven additional countries with data for domestic (but not cross-border) letter volumes, we are able to construct our own estimate of cross-border volumes by using a reasonable ratio of cross-border to domestic letter mail volumes.¹⁵ This results in a data set of inbound and outbound volumes (in number of items) for 182 countries and territories.

As a second step, we set the estimates of inbound and outbound flows in 2011 for each designated postal operator relative to the 2011 inter-regional mail flows to and from for the region where the postal operator is located.¹⁶ This provides us with a ratio (for each designated operator) of (i) the inbound flow relative to the total inbound flow to the region and (ii) the outbound flow relative to the total outbound flow from the region.

As a third step, we make an important assumption about proportional regional participation, where we depart from the ratios from step two and apply them to inter-regional mail flows from 2014. By applying this assumption, we are able to estimate data points for each bilateral mail flow as a percentage of the global cross-border mail flow in 2014. While this methodology fails to take into account factors such as distance and international relations,¹⁷ it is the best available approximation of bilateral flows and does not suffer from further data gaps.

Assumption of proportional regional participation

The assumption about proportional regional participation allows us to estimate the percentage of the world’s mail flow that goes from a country i in region A to a country j in region B based on the following information:

• (i) The percentage of region A’s outbound flow originates in country i,

• (ii) The percentage of region B’s inbound letter mail flow that is delivered to country j

• (iii) The share of the global total cross border mail flow that goes from region A to region B

For each mail stream that goes between countries from different regions, we apply the following calculation:

$$ X_{ij} = \frac{{O_{i} }}{{O_{A} }} \times \frac{{I_{j} }}{{I_{B} }} \times X_{AB} $$

(1)

For each mail stream that goes between countries within the same region, we apply the following calculation:

$$ X_{ij} = \frac{{O_{i} }}{{O_{A} }} \times \frac{{I_{j} }}{{I_{A} \times (1 - \left( {\frac{{I_{i} }}{{I_{A} }}} \right))}} \times X_{AA} $$

(2)

$X_{ij}$ :

percentage of world’s mail flow that goes from country i to country j

$X_{AB}$ :

percentage of world’s mail flow that goes from region A to region B

$X_{AA}$ :

percentage of world’s mail flow that goes within region A

$O_{i}$ :

total outbound mail flow from country i

$O_{A}$ :

total outbound mail flow from region A

$I_{j}$ :

total inbound mail flow to country j

$I_{B}$ :

total inbound mail flow to region B

The assumption in (1) and (2) is essentially the same but the calculations differ in the relative inbound. The difference is due to the fact that the total inbound flow to region A also includes mail destined for country i. This is a pragmatic solution to solve a mechanical problem in the model because we do not want to include domestic mail.

As a fourth step, we convert each bilateral flow (until now measured as a percentage of the global flow of international mail) into an estimate in terms of volumes (number of items). We do this by multiplying it with the global volume of international mail.¹⁸

As a fifth step, we split each bilateral mail flow according to the different letter formats (P, G, and E). In order to do this, we use a UPU (2014d) survey among 49 designated postal operators that contain regional estimates for the distribution of formats (in percentage of number of items and weight). This study also provides estimates of the average weights per format. By assuming that the distribution of P, G and E items in a specific bilateral mail flow between two regions is the same as the distribution of P, G, and E items in the total inter-regional mail flow between the same two regions, this provides us with a per-format estimate of letter mail volumes for each bilateral mail stream.

As a sixth step, we apply the average weight for each product type (from the UPU survey referred to above) in order to get the bilateral flows measured in kilograms instead of number of items.

Based on this, we are able to estimate bilateral cross-border flows of letter post items between 183 countries and territories worldwide. Depending on the size and structure of bilateral mail flows, the current design of the terminal dues system will affect designated operators differently. For example, large outbound flows of cross-border mail may imply that designated postal operators experience large positive transfers on the outbound side. Similarly, in combination with a negative difference between actual terminal dues received and the equivalent domestic postage rate, large inbound may imply that designated postal operators experience large negative transfers on the inbound side. Understanding the structure of bilateral mail flows will thus be very helpful when trying to understand the structure of net financial transfers created by the current terminal due system.

Lastly, for the new scenario of growing e-commerce we select the set of exporting countries, namely China, Hong Kong, India, Singapore and Malaysia. We then scale up all the flows of the E-format going from these countries to any EU-28 or OECD country by an annual growth rate of 50 % from 2014 to 2018 (United States Postal Service Office of Inspector General 2014a).

Terminal dues rates

Actual terminal dues rates are directly based on the UPU method set out in the 25th UPU congress for the period of 2014 through 2017.¹⁹ For the proposed scenario, we base our terminal dues rates on the proposal for the UPU terminal dues system for the 2018–2021 cycle, as outlined by the UPU (see UPU 2016a). In the target system the terminal dues charges consists of a per-item rate and a per-kilogram rate. Before the application of any caps and floors the rates are calculated based on two reference tariffs, a 20 g small letter and a 175 g large letter.²⁰ A linear relationship between these two are drawn and evaluated at the weight of an average item. If the revenue from an average package are above the cap revenue, the cap rates are applied. Equivalently, if it is below the floor revenue, the floor rates are applied. If the average revenue is between the cap and floor a target rate is applied.

The target rates per kilogram are calculated as:

$$ R_{w} = 70\;\% \times \frac{{M \times \left( {W_{avg} - 0.01} \right) + DP_{1} }}{{R_{wfl} \times W_{avg} + R_{Ifl} }} \times R_{wfl} $$

(3)

and the target rates per item are calculated as:

$$ R_{I} = 70\;\% \times \frac{{M \times \left( {W_{avg} - 0.01} \right) + DP_{1} }}{{R_{wfl} \times W_{avg} + R_{Ifl} }} \times R_{Ifl} $$

(4)

where,

M :

Constant rate of change = (DP2–DP1)/(0.175–0.01)

DP ₁ and DP ₂ :

Domestic postage rates without VAT for 0–20 g P and 100–250 g G

W _avg :

The average weight of an letter post item, set to 81 g in the current system

R _wfl :

Floor rate per kilogram

R _Ifl :

Floor rate per item

The methodology of the proposed regulatory system is a continuation of the current system. The main difference is a separation of the E-format letters. Whereas the P and G format can be calculated in the same way as before, the E letters now get slightly different parameters. The revenue of a letter of average weight, i.e. the numerator in Eqs. (3) and (4), is based on the average weight of an E-letter letter (375 g) and the item and weight rates from the current system. The cap and floor rates are also different for the E-format. In particular, the ratio between the per-item floor rate and the per-kilo floor rate (this is the so-called item-to-kilo ratio) is higher. Although often expressed as a percentage by the UPU, it is in practice better described as a weight. It is the weight for which, an item of a specific weight and at a specific kilogram rate, is equivalent to the item charge. A higher item-to-kilo means the structure of the rates is “flatter”, something that is intended to better reflect the price structure of the small packets (UPU 2016a) (Table 1).

Table 1 Parameters for the two cycles

Counterfactual terminal dues rates

As a proxy for the counterfactual terminal dues rates, we depart from the domestic (end-to-end) postage rates for single-piece items of three different formats (P, G, E) and weights. This means that we use more granular data than what the UPU uses to calculate terminal dues. In particular, the terminal dues does not use prices for items of E format. In order to reflect the price charged for delivery of domestic letter mail, we apply an adjustment factor of 70 %.²¹

Instead of making a simple linear relationship between two tariffs, we use information about different weight-steps for three products. Our methodology uses three weight steps for small letters, five weight steps for large letters and seven weight steps for small packets. This allows us to calculate equivalent domestic postage rates that mirror the actual situation better than what would have been the case if only two prices would have been used. We collapse the data on domestic postage rates to a level that fits our counterfactual purposes. In order to calculate an average postage rate per item, we thus need to make assumptions about the distribution of weights. These assumptions are primarily based on a UPU study (see UPU 2014d) containing estimated distributions on different weight steps for each product type P, G and E. Since we also have information of average weights of mail flows we can vary the distributions based on average weight. Once the weight distribution for each format is established, it is straightforward to calculate a domestic rate per item for each product type.

For industrialized countries and all European countries, data on domestic postage rates mainly comes from a survey conducted by WIK consulting, containing a comprehensive overview of domestic postage rates in 2013 (see WIK 2013). For other countries domestic postage rates are collected from a UPU database with 2008 priority domestic rates by product type and weight step.²² These rates are adjusted based on inflation to create an estimate of tariff levels in 2014 and subsequently 2018. Our data set contains domestic postage rates for 164 countries and territories. Combined with the data set of bilateral mail flows (consisting of 182 countries and territories), this leaves us with a set of 155 designated operators (for which we have estimates of all necessary parameters) to include in the analysis.

Calculations for net transfers

For each bilateral country pair we estimate:

$$ \begin{aligned} \pi_{ij}^{I} & = X_{ji} (TD_{ij} - EDP_{ij} ) \\ \pi_{ij}^{O} & = X_{ij} (TD_{ji} - EDP_{ji} ) \\ NT_{i} & = \pi_{ij}^{I} - \pi_{ij}^{O} \\ \end{aligned} $$

where $\pi_{ij}^{I}$ is country i’s loss (gain) on inbound letter mail from country j expressed as the difference between what country i gets in terms of terminal dues from country j today and what it would get in the counterfactual scenario. $X_{ji}$ is the letter mail flow subject to terminal dues going from country j to country i, $TD_{ij}$ is the actual terminal dues rate that country i receives as revenue from country j and $EDP_{ij}$ is the equivalent domestic postage (i.e. the counterfactual terminal dues rate).

Equivalently, $\pi_{ij}^{O}$ is country i’s gain (loss) on outbound letter mail to country j expressed as the difference between what country i pays in terms of terminal dues to country j today and what it would pay in the counterfactual scenario.

$NT_{i}$ is thus the net transfer for country i related to its bilateral cross border exchange of mail with country j. In order to get the total net transfer for country i, one has to conduct the same analysis for all bilateral mail flows to and from country i.