Do cross‐border patents promote trade?

While we would expect that cross-border patents are used to protect a technology is made available in another country, could either be produced locally or imported. International patent filings therefore complements substitutes international trade. This study combines data on patenting and trade for 149 countries 249 industries between 1974 2006 with “three-way” panel model addresses several biases emphasized the literature order provide systematic analysis of how bilateral responds filings. We find have positive (complementary) overall effect patent-filing country's exports patent-granting country no imports flowing opposite direction. These effects vary substantially across industry groups, promoting significantly more export growth high demand elasticity relatively downstream supply chains. also patents, once obtained, associated increased even jurisdictions weak intellectual property regimes. Les brevets transfrontaliers favorisent-ils le commerce international? sont généralement utilisés pour protéger une technologie qui est rendue disponible dans un autre pays, mais cette pourrait y être produite localement ou importée. dépôts de à l'étranger et pourraient donc complémentaires substituables. Cette étude combine des données sur les couvrant pays entre 2006. Nous utilisons modèle effets fixes « trois voies » répond plusieurs biais soulignés la littérature afin fournir analyse systématique l'impact international. constatons que ont effet global positif (complémentaire) exportations depuis résidence l'inventeur vers délivre brevet, aucun importations circulant direction opposée. Ces varient considérablement d'une industrie l'autre: favorisent croissance significativement plus élevée lesquelles l'élasticité demande forte, ainsi relativement en aval chaînes d'approvisionnement. également brevets, fois obtenus, associés augmentation du commerce, même juridictions où protection propriété intellectuelle faible. As technological advancements been driving force behind economic globalization, flows mirrored one ways hardly seem coincidental. Both grown at rapid paces relative world GDP since 1970s (figure 1). And non-OECD become involved both as recipients originators new over last decades 2), just they high-tech manufactured goods this same time frame. In policy circles, lack rights widely cited an important barrier trade, provisions protections increasingly central agreements intended promote regulate (Maskus Penubarti 1995, McCalman 2001, Limão 2007, Saggi 2016). Growth flows, 1972–2008 NOTES: The left graph shows total share patents. right 1972 2008 versus real GDP. All values normalized 1 1972. World deflated using US deflator. Patent counts raw number filed given year SOURCES: PATSTAT, WDI Counts by origin type destination type, NOTE: year. include only. SOURCE: PATSTAT yet, despite these suggestive anecdotes, much still remains unknown about what extent influence Through lens standard theory, generally firms file foreign market inherent product intend sell market. But, empirically, do not evidence effective regard, when, nor can confidently say whether observe truly reflect intent opposed other motivations, such “strategic patenting,” which aims instead establish assets use potential future litigation. Indeed, when considered carefully, theory does give unambiguous guidance should related positively negatively another. To incentivizes diffusion, example, might increase activity coincide reduced exports. Firms producing there directly, passing increasing alternate modes sales FDI licensing.11 Other reasons why relationship theoretically ambiguous discussed following section. depend factors and/or markets. For stronger strong (IPR) than only IPR protections. some others. Moreover, though vibrant empirical has investigated national systems, country-level indicator, few studies examined affected actual origin, specific represent. Notably, none trends mentioned above setting large disaggregated industries. conspicuous gap possibility may explaining evolution industry-level patterns well understanding importance systems granular level. general, promoted impeded it largely open question implications policy. deliver answers, assemble extensive database 32-year period from then implement rigorous gravity estimation methodology specifically assesses increases followed industry. By pooling our study, able construct econometric averages had each investigate heterogeneous different types exploit structure demonstrate robustness findings natural concerns reverse causality patenting. bringing together way, unpack facts conjunction full, unconditional sample, complementary trade: average tends stimulate direction, significant seen economically significant: difference “high value” flow “low patents—lying respectively log-deviation below log-mean distribution—is 8 . 87 % , half free agreement (FTA) trade.22 results, explained section 5, evaluated peak patent's life. calculate stocks based value-weighted weighted family size. Digging deeper, sources heterogeneity interest. examine differentiated where monopoly product's distinguishing characteristics valuable. Contrary expect, actually overwhelming opposite: larger less valuable protecting but serve cost quality advantages producers competing products. test trade-promoting institutions. Interestingly, any strength regime indicating affect independently regime. Studying co-evolution presents challenges. part because complex objects work empirically: “lumpy” (occurring irregular points time, long durations) highly derive in. Furthermore, documented literature, accounting many dimensions naturally arise multilateral context quickly becomes computationally difficult estimations aggregate (see, e.g., Glick Rose 2016, Larch al. 2019). resolve various challenges, draw recent methodological innovations complement focus First, lumpiness bilateral, industry-specific stock reflecting gradual obsolescence old addition lag diffusion (Popp 2002, Popp 2003). Our accounts lasting previously Second, address issue others, weight their Third, specification specificity effectively estimations, its own set exporter–industry–time, importer–industry–time country-pair–industry fixed effects. numerous bias shown similarly show assessing patents.33 was first Baier Bergstrand (2007) Pooled estimators literature. French (2019) describes detail pooled described paper. work, Schmidt Steingress (2018) Kinzius similar specifications harmonization non-tariff barriers, respectively. Finally, capitalize Correia (2020) (2021) estimate preferred Poisson pseudo-maximum likelihood (PPML). PPML ensures estimates consistent subject aggregation (French above, prior examining limited scope. Most papers area constrained single (either exports), relationship, relationships (Anderton 1999, Greenhalgh 1990, Antimiani Costantini 2013, Brunel 2019, Chen Chalioti 2020). exception Palangkaraya (2017), who successful applications Using 13 host markets, 28 exporting countries, find, do, exports.44 precise, (2017) presence apparent application process “blocking art” importing deters investigation goes beyond ways. wider range effects, allowing directions key cross-sectional variation outcomes invention offices processes, whereas longitudinal changes within pair markets result approach advantage explicitly controlling omitted cause propensity particular correlated country. facilitates expansive varied data, years, (especially) variety OECD destinations. sum, contribute greatly expanding geographical scope existing analyses matches detailed level disaggregation. notably dimension control unobservable pairs simultaneously well-known articulated cannot addressed firm-level data. Because surmount pursue comprehensive isolate relationship. amount devoted active looked “gravity”-like determine distribution knowledge capital flows. Keller Yeaple's (2013) influential “gravity knowledge” Figueiredo al.'s (2015) industry- location-specific clustering citations. Earlier contributions Robbins (2006) Scherngell Hu (2011). (2004) reviews earlier being channels. theoretical meanwhile, very rich. observed offer consensus view substitutes. Lai Qiu (2003) Grossman (2002), South decrease stimulating innovation-intensive production thereby reducing North–South comparative advantage. Yang Maskus (2009) that, innovating North via exports, licensing, depending firms' optimal mode strategic incentives transfer know-how South.55 ambiguity (2001). Bond (2014) regimes threat compulsory licensing induce firm directly rather voluntarily license patented product.66 Of course, debates surrounding TRIPS issues stimulated studying topics; surveys, see (2016) (2018). addition, likely capture cases concurrent decisions (or, least, firms). notes, “in typical year,” 80% royalty payments parent subsidiaries. thus broadly relates role play determining intermediate inputs firm's chain organized among entities countries. Representative works Nunn Trefler (2013), Antràs Chor Alfaro (2019). Also Blonigen (2001), finds final industries, part, goods; however, driven solely consumption goods. Looking position textured (i.e., closer demand) those upstream. while sparse, body relating institutions See, (1995), Briggs Park (2014), Boring (2015), Awokuse Yin (2010) Ivus (2010). (2002) provides survey area. design individual line research. varies regime, unique information captured indicators. noted introduction, analyzing requires ambitious framework. many-country destination, time—to denoted i, j, k t, aim identify exploiting origin–destination–industry (or “pair–industry”) Part makes task interconnected nature known least Anderson van Wincoop (2003), correctly determinants two recognizing potentially confounding played size geography respect third countries.77 (and 2003), function “multilateral resistance” faces network, controls observables traditional wholly inadequate isolating linkages. sizes change resistance themselves source accounted (Baldwin Taglioni 2007). A further challenge technologies from. It necessarily case, aerospace equipment will cosmetics require complexity patent, object interest industry, time. Setting aside raised regarding log-linear models, rich place (2) leading challenges commonly estimation. need account interdependencies (industry-level δ i t ψ j ) (captured η ), costs currently pair.1212 observation time-varying variables tend biased absence credited (2007). More generally, note appealing logic identification speaks objectives. affects completely stripping away industry–country-specific fluctuations patenting, country–industry–time allow us pair–industry β through pairs, analysis.1313 uses methodology, possible fourth indexed common all pair. prefer stick three-way conforms intuition results estimations. document 5.3, materially adding effect; inference comparison. intuitive terms, ensure picks up value market, innovation patent. That say, “direct” irrespective filing—by productivity inventing firm, example—we destinations, ones owns direct controlled exporter–industry–time sure, inventions aimed case However, varieties perceived substitutable. support hypothesis, country-specific results. Nonetheless, clear interpretation depends assumption reflects protection, itself protected. discuss next section, informs decision size, take measure innovation. suggests ambiguous. fix ideas, reasonable start supposition exporters enjoy Importantly, baseline hypothesis—that exports—but guides possibly main contrast primarily varieties, perspective implies superior greater higher substitution replication patent-holder's represents profitability and, similarly, assertion exclusive disputed impact rationales, strongly high-elasticity low-elasticity industries.1414 taking lower enhance quality. differentiate eyes end-users. pronounced find. consider interactions trading obtaining enforcing indicators relationships, log distance sharing legal system language, dummy (“medium” “high,” “low” category). inclusion levels helps link often found 2010), interesting wonder vehicle effect. One remaining lies endogeneity if (partner-specific) import subsequently partner Import competition, stifle home (Autor Batrakova Dechezleprêtre 2013). Alternatively, technology-intensive point innovative or, contrary, sufficient plateaued. multinational corporations influenced locate production. precursor corporation filing setup capacity abroad. concerns, ideally want instrument time-varying, pair-specific s