Where should Germany sell HS 870323?

For an exporter of vehicles: spark — ignition internal combustion reciprocating piston engine, cylinder capacity exceeding 1500cc but, the market-entry scorecard ranks destinations on a composite of demand size, demand growth, tariff exposure, ease of doing business, and a preference-fit term that combines the destination’s economic complexity with the product’s complexity. The top three picks in 2024 are Türkiye, Georgia, Russian Federation, in that order.

How the score is built

The entry score is a product of five normalised factors, each on the unit interval so the composite reads as a pseudo-probability of a good fit:

entry_score = demand_size × demand_growth × tariff_advantage × ease_of_business × preference_fit

demand_size       = log10(imports_usd + 1) / log10(max_imports + 1)
demand_growth     = clip( (CAGR_2019_2024 + 0.15) / 0.45 , 0 , 1 )
tariff_advantage  = 1 / (1 + simple_avg_tariff / 10)
ease_of_business  = World Bank Doing Business score (2020) / 100
preference_fit    = sigmoid( (ECI_dest + PCI_product) / 4 )

Demand size is the destination’s total imports of this HS6 in 2024(CEPII BACI, values × 1000 to get USD; Gaulier & Zignago 2010). The bilateral gravity workhorse (Anderson & van Wincoop 2003, AER) motivates the tariff-advantage term: ad-valorem trade resistance enters multiplicatively in the import-price wedge, so a 10% applied tariff roughly halves the attractiveness of the market relative to a zero-tariff alternative. The growth term proxies the extensive-margin pull documented by Helpman, Melitz & Rubinstein (2008, QJE), where positive-market probability responds to the expected value of entry. Preference fit follows Corcos, Del Gatto, Mion & Ottaviano (2012, Economica) on firm heterogeneity and the complementarity between exporter capability and destination sophistication: the sigmoid of (ECI_dest+ PCI_product) centres the term at 0.5 when the sum is zero, and saturates at the tails. ECI is the Atlas of Economic Complexity measure (Hidalgo & Hausmann 2009,PNAS; Hausmann et al. 2014 Atlas).

The world picture

Figure 1 shows the composite score across all 223 destinations that imported HS 870323 in 2024. Darker shades mark higher composite scores, i.e. markets where size, growth, tariff access, regulatory quality, and preference fit coincide. The exporter country itself (DEU) is excluded from the ranking.

What drives each top-15 destination

Figure 2 breaks out the five factor values for each of the top 15 destinations by composite score. Each factor is unit-interval normalised, so the bar lengths are directly comparable within a row-group, and the 75 bars together show which component is pulling each market up or pulling it down. A destination with a tall tariff advantage bar and a short demand size bar is a low-barrier but small market; the reverse is a big-market hurdle.

Size vs growth: big safe bets vs small rising bets

Figure 3 plots destinations on log-import-size against 5-year import CAGR. The upper-right quadrant is the big-and-growing demand: canonical priority markets for a scale-oriented exporter. The upper-left is small but rising, where first-mover entry yields a larger future share; the extensive-margin logic in Helpman, Melitz & Rubinstein (2008) applies most cleanly here. The lower half is shrinking demand, where entry costs are rarely worth the falling stream of revenue.

Tariff landscape for the top destinations

The CEPII pref_tariff_hs6 table aggregates WITS-TRAINS applied duties by reporter, partner, and HS6. Since this workbench stores reporter-partner level detail but not a clean MFN-vs-preferential flag at the aggregate row, the table below reports the simple-average applied rate across all partners observed for the destination in 2023(a proxy for typical trade-weighted exposure), alongside the minimum applied rate (which approximates the most-favourable preferential rate) and the maximum (which approximates the MFN ceiling). Where data is not available, the row is marked.

The penetration plane: preference fit vs tariff advantage

Melitz (2003, Econometrica 71(6): 1695–1725) models the exporter’s destination choice as selection above a productivity threshold that depends on destination-specific fixed and variable trade costs. Arkolakis (2010, JPE 118(6): 1151–1199) adds an endogenous within-destination marketing cost that produces the long tail of small exporters observed in customs data. Eaton, Eslava, Kugler & Tybout (2008, in Helpman et al., The Organization of Firms in a Global Economy) document exactly this pattern for Colombian exporters: entry is lumpy, survival is short, and penetration rises slowly with destination size and capability match. The plane below plots two model-consistent axes: tariff advantage (variable trade cost wedge; Melitz’s τ) and preference fit (destination–product capability match; Arkolakis’ penetration shifter). The upper-right quadrant is where both margins favour entry, matching the “export superstar” destinations of Eaton et al.

Total entry cost: tariff, NTMs, and distance stacked

The composite in Figures 1–2 normalises each factor to a unit-interval score, which helps ranking but obscures magnitudes. Figure 6 returns to raw ad-valorem-equivalent percentage points for the three variable-cost wedges that any exporter of HS 870323 faces at the border. Tariffs enter as WITS-TRAINS simple averages (%). NTMs enter as Kee, Nicita & Olarreaga (2009, Economic Journal 119(534): 172–199) tariff-equivalents, scaled at a rule-of-thumb 0.5 pp per UNCTAD MAST NTM count for the destination at this HS6 (they estimate per-HS6 coefficients; the mean of their cross-country distribution is close to this magnitude for non-agricultural goods). Distance enters at Head & Mayer (2014, Handbook of International Economicsvol. 4) gravity elasticity of roughly −1 on log bilateral distance, converted to an ad-valorem-equivalent iceberg cost at trade elasticity σ = 5 (Anderson & van Wincoop 2003, AER 93(1)), giving 5 pp per log-point of distance relative to a 1000 km baseline. Sums below 1000 km or below-baseline rows are clipped at zero so the distance term only penalises, never rewards.

Partial-effects summary: which barrier dominates on average?

Figure 6 shows the full tariff-plus-NTM-plus-distance stack per destination; a single policy-relevant question cuts across the destination detail — on average across the top-50 destinations for HS 870323 from DEU, which wedge contributes the most ad-valorem-equivalent percentage points to the entry hurdle? Anderson & van Wincoop (2003, AER) and Head & Mayer (2014, Handbook vol. 4) treat the three wedges as additive in log-iceberg cost; under that approximation the cross-destination mean pp per factor is the partial-effect estimate relevant to bilateral trade-cost policy. Kee, Nicita & Olarreaga (2009, EJ) and Hummels & Schaur (2013, AER103(7): 2935–2959) emphasise that NTMs can exceed tariffs in ad-valorem-equivalent magnitude for differentiated goods; the single-HS6 view below tests that claim for this specific product.

How concentrated is global demand for this HS6?

The composite ranking and decomposition treat every destination as a candidate market, but the empirical question for an exporter sizing sales-team coverage is: how many destinations actually account for most of world demand at this HS6? A Lorenz curve on import value answers this directly. The cumulative-share curve is constructed by ranking destinations from largest to smallest importer and plotting the cumulative share of world imports against the cumulative share of destinations covered. A curve that hugs the top-left corner means a handful of markets carry global demand (concentrated, network-style export industries: Eaton & Kortum 2002, Econometrica70(5): 1741–1779); a curve close to the 45-degree line means demand is broadly diffused (Krugman 1980 monopolistic-competition intuition extended to many small markets).

Synthesis

The five figures read as a Melitz–Arkolakis destination-ranking pipeline applied to a single HS6. Figure 1 maps the composite score across the world, Figure 2 decomposes the top 15 into the five normalised factors, Figure 3 separates the size–growth tradeoff that anchors prioritisation, Figure 4 opens the tariff schedule underlying the tariff-advantage term, and Figure 5 plots the cost-and-fit plane that governs firm-level entry in Melitz (2003) and Arkolakis (2010). The score is a ranking aid, not a forecast: Eaton, Eslava, Kugler & Tybout (2008) document that even with a favourable destination ranking, the modal Colombian exporter exits within two years, so destination selection is necessary but not sufficient for sustained entry. The complementary question — which firms within the exporter country can actually cross the productivity cutoff into each destination — requires firm-level customs data and is outside the workbench universe.

Data gaps and caveats

• World Bank Doing Business vintage. The workbench parquet has Doing Business data through 2020 only; the programme was discontinued in 2021 (World Bank 2021 data-irregularities review). For a current-year fit, substitute the successor Business Ready (B-READY) series when it is added to the workbench, or proxy via WDI indicators (e.g. NY.GDP.PCAP.CD for wealth, IC.LGL.CRED.XQ for credit depth). TODO: add B-READY once CEPII/WB release a clean HS-matched panel.
• Tariff coverage. Of the 223 destinations with imports of this HS6, 83 have a WITS-TRAINS applied-tariff record in 2023; the other 140 use the median of 0.4% as a scoring fallback. Preferential-vs-MFN distinction at the row level is not clean in the current parquet, so avg/min/max of partner-level simple-averages is reported.
• RTA matching. The workbench rta module exists but is not joined here: the preference-fit term uses ECI as a stable proxy rather than exporter-destination RTA flags. A fuller build would multiply the tariff-advantage factor by an RTA indicator for the DEU-destination pair, reducing the effective rate toward the minimum observed.
• ECI coverage. 223 of 223destinations have an ECI record in 2024; missing values default to 0 (world median), which biases the preference-fit term toward 0.5 for small or poorly-covered destinations.
• Scoring is a ranking aid, not a forecast. It does not model entry cost, distribution capacity, or competitive rivalry at the destination; Corcos et al. (2012,Economica) show that within-destination firm heterogeneity matters more for entry success than country-level averages.

How teams use this

• Export-development prioritisation. Trade-promotion agencies can use Figure 1 to pick the three to five destinations where a sectoral promotion budget has the highest expected return.
• Distributor search. Figure 2’s factor decomposition identifies the binding constraint per destination (tariff? size? regulatory?) so the business-development playbook is pre-scoped before shortlisting agents.
• Trade-policy lobbying. Large-demand destinations with low tariff advantage (left-hand bars in Figure 2) are candidates for RTA/FTA priority lists; the gravity-theoretic gain from removing the tariff wedge is large when size × growth is already high.

References

• Anderson, J. E., & van Wincoop, E. (2003). “Gravity with Gravitas: A Solution to the Border Puzzle.” American Economic Review 93(1): 170–192.
• Arkolakis, C. (2010). “Market Penetration Costs and the New Consumers Margin in International Trade.” Journal of Political Economy 118(6): 1151–1199.
• Corcos, G., Del Gatto, M., Mion, G., & Ottaviano, G. I. P. (2012). “Productivity and Firm Selection: Quantifying the ‘New’ Gains from Trade.” Economic Journal 122(561): 754–798.
• Eaton, J., Eslava, M., Kugler, M., & Tybout, J. (2008). “The Margins of Entry into Export Markets: Evidence from Colombia.” In E. Helpman, D. Marin, & T. Verdier (eds.), The Organization of Firms in a Global Economy, Harvard University Press.
• Gaulier, G., & Zignago, S. (2010). “BACI: International Trade Database at the Product-Level. The 1994–2007 Version.” CEPII Working Paper 2010-23.
• Hidalgo, C. A., & Hausmann, R. (2009). “The Building Blocks of Economic Complexity.” PNAS 106(26): 10570–10575.
• Hausmann, R., Hidalgo, C. A., Bustos, S., Coscia, M., Simoes, A., & Yildirim, M. A. (2014). The Atlas of Economic Complexity: Mapping Paths to Prosperity. MIT Press.
• Helpman, E., Melitz, M., & Rubinstein, Y. (2008). “Estimating Trade Flows: Trading Partners and Trading Volumes.” Quarterly Journal of Economics 123(2): 441–487.
• Melitz, M. J. (2003). “The Impact of Trade on Intra-Industry Reallocations and Aggregate Industry Productivity.” Econometrica 71(6): 1695–1725.
• World Bank (2020). Doing Business 2020. Washington, DC: World Bank.