Product space: A tool to identify where to look, not what to do

tepav
The Economic Policy Research Foundation of Turkey
Product space: A tool to identify
where to look, not what to do
Examples & Applications from TEPAV Projects
Istanbul, 28 January 2014
Agenda
 A new tool to analyze why some countries
are rich and others are poor
 What does product space tell about the
evolution of Turkey’s productive structure?
 Three examples from TEPAV’s project
portfolio that we utilized product space
methodology
The hidden structure in the
comparative advantage of nations
The hidden structure in the
comparative advantage of nations
What does this pattern mean?
 Some countries make few things, while others
make many things
 Some products are made by many countries,
other products are made by few countries
 Countries that make few products, make
products made by many countries
 Countries that make many products, also make
products made by few other countries
 The world does not specialize
Countries
Products
Diversification
Ubiquity
Degree (Countries)
Degree (Products)
kc   M cp
p
kc1=3
kc2=4
k p   M cp
c
Product p1
kp1=2
Product p2
kp2=2
Country C1
Country C2
Product p3
kc3=1
kp3=1
Country C3
Product p4
Hidalgo CA, Hausmann R Proc. Natl. Acad Sci. (2009) 106(26):10570-10575
kp4=3
Countries
Products
Countries
Capabilities
Products
Intuition
 Countries differ in the diversity of their capabilities
and products require different sets of capabilities
 Countries that have more capabilities will be able
to make more products
They would be more diversified
 Products that require more capabilities will be
made by fewer countries
Products will be less ubiquitous
Intuition (cont’d)
 Countries that have more capabilities will be able
to make products that require more capabilities
i.e. products that are less ubiquitous
 Hence, countries that have more capabilities will
be more diversified and will make less ubiquitous
products
 Diversification of countries and ubiquity of
products are negatively correlated
They are indirect measures of the capability set
of countries
Diversification and ubiquity - 2009
Economic complexity explains
income per capita
Deviations explain future growth
Complexity in 1985, controlling for initial GDP per capita
Hidalgo CA, Hausmann R Proc. Natl. Acad Sci. (2009) 106(26):10570-10575
Let us look inside Turkey
(Cities) Sorted by Diversification
10
20
30
40
50
60
70
80
200
Source: data from TOBB
400
600
800 1000 1200 1400 1600 1800 2000
(Products) Sorted
Products:
sortedbybyUbiquity
ubiquity
Diversity vs. ubiquity
60
ARDAHAN
TUNCELİ
50
Diversity and average ubiquity of Turkey's cities
BİNGÖL
IĞDIR
HAKKARİ
BAYBURT
AĞRI
40
GÜMÜŞHANE
MUŞ
SİİRT
BİTLİS
ERZİNCAN
20
30
RİZE
NİĞDE
SİNOP
KARS
ŞIRNAK
ARTVİN
BARTIN
BATMAN
ADIYAMAN
KASTAMONU
ERZURUM
VAN
KARABÜK
AKSARAY
KIRŞEHİR
OSMANİYE
AMASYA
GİRESUN
KİLİS
KARAMAN
DİYARBAKIR
ELAZIĞ
ŞANLIURFA
MARDİN
NEVŞEHİR
SİVAS
TOKAT
BURDUR
ORDU
DÜZCE
YOZGAT
ÇANAKKALE
MUĞLA
MALATYA
ZONGULDAK
ÇANKIRI
ISPARTA
EDİRNE
BOLU
KAHRAMANMARAŞ
AFYON
KIRIKKALE
UŞAK
TRABZON
KÜTAHYA
HATAY
YALOVA
KIRKLARELİ
AYDIN
ÇORUM
ANTALYA
DENİZLİ
SAMSUN
İÇEL
BİLECİK
SAKARYA
BALIKESİR
ESKİŞEHİR
KAYSERİ
MANİSA
GAZİANTEP
ADANA
TEKİRDAĞ
KONYA
BURSA
KOCAELİ
ANKARAİZMİR
10
İSTANBUL
0
500
Source: Turkstat, Tepav Calculations
1000
k0 (diversification)
1500
Diversity explains GDP per capita
differences within Turkey
8.5
KOCAELİ
8
BOLU
6.5
7
7.5
YALOVA
İSTANBUL
KIRKLARELİ
İZMİR
MUĞLA
ANKARA
ZONGULDAK
EDİRNE
BİLECİK
ÇANAKKALE
BURSA
KIRIKKALE
TEKİRDAĞ
ESKİŞEHİR
İÇEL
MANİSA
ADANA
NEVŞEHİR
SAKARYA
ANTALYA
ARTVİN
BALIKESİR
KARAMAN AYDIN
DENİZLİ
BURDUR
NİĞDE
RİZEKASTAMONUHATAY
KİLİS
SAMSUN
ÇORUM
KAYSERİ
KONYA
KARABÜKELAZIĞ
KÜTAHYA
ISPARTA
GAZİANTEP
AMASYA
UŞAK
TUNCELİ
KAHRAMANMARAŞ
KIRŞEHİRMALATYA
TRABZON
SİNOP
GİRESUN
TOKAT
SİVAS
AFYON
DİYARBAKIR
DÜZCE
ÇANKIRI
OSMANİYE
ERZİNCAN
BATMAN
GÜMÜŞHANE
ERZURUM
AKSARAY
SİİRT
ORDU
BARTIN
BAYBURT
ŞANLIURFA
ADIYAMAN
YOZGAT
IĞDIR HAKKARİ
KARS
MARDİN
VAN
BİNGÖL
ARDAHAN
BİTLİS
ŞIRNAKAĞRI
MUŞ
6
Income per capita (logs)
ln(gdppp) (2000)
Diversity vs GDP per capita
2
4
6
Diversification (ln(k0))
Fitted values
lngelir
Diversification
Source: Turkstat, Tepav Calculations
8
What makes growth difficult?
 The chicken and egg problem
 You cannot make new products because you lack
the capabilities
 You don’t want to accumulate the capabilities
because the products that need them are not
being made
Because of other missing capabilities
 How does the world deal with this?
 By moving towards “nearby” products
Four elements of strategy
 How near is the product?
How easy would it be to become good at it?
 How large is the global market in the product?
How big is the potential gain?
 How sophisticated is the product?
How big is the wage advantage with competitors?
How profitable would it be if one succeeds to make
it?
 How strategic is the product?
How will it improve my position in the product
space, putting me closer to other products?
What is in the nearby forest?
prody
16000
18000
14000
the
is20000
How sophisticated
22000
product?
Industrial,
laboratory
furnaces, ovens,
incinerators
Taps, cocks,
valves for pipes,
tanks, boilers,
etc
Medicaments,
therapeutic,
prophylactic
use, n dosage
Lifting, handling, loading
Parts for
machinery nes.
use with
lifting,
moving
machinery
Plastic plate, sheet, film,
foil, strip, cellular, ne
14000
16000
18000
mhopen_forestmin
20000
How strategic is the product?
Bubbles are proportional to world trade in that product!
Which industries in which Turkish cities are
the most likely to emerge?
Konya: Aluminum radiator
1
Kayseri:
Food
processing
machinery
Mus: PVC
joinery
0.9
10
0.8
Cities (sorted by diversification)
Adana: Lift
components’
parts
20
0.7
30
0.6
40
0.5
50
0.4
0.3
60
0.2
70
0.1
Ardahan:
Ready-mixed
concrete
80
500
1000
1500
Products (sorted by ubiquity)
2000
Recap and the way forward…
 Turkey has a great position in the “Product Space”




and has moved quickly in it
It has very large internal differences across its
many cities
Diversification is the road forward at the national
and city level
This requires addressing the “chicken and egg
problems”
The quality of public-private collaboration in
identifying opportunities and obstacles is the true
foundation of sustainable competitiveness
Four examples from TEPAV’s work
on competitiveness
1. Identification of low hanging fruits for Turkey’s
cities (with Ministry of Economy)
2. Diversification strategy design for Zafer Region
(with Zafer Dev. Agency)
3. Increasing the competitiveness of Turkish
machinery industry (with Ministry of Economy)
4. Tool kit for informing regional competitiveness
agenda (with UNDP & GAP Regional Agency)
What new industries could emerge
in Turkey’s cities?
 We sought answers to this question for the
Ministry of Economy
 The project was an application of product space
methodology with one major modification
 World  Turkey, Countries  Turkey’s cities
 Productive structures of each city were put onto
the product space and low hanging fruits for each
city were estimated
Three cities that have different levels of
development on product space
Ankara
Denizli
Diyarbakır
Parts and accesories
of weapons
Bed, table, kitchen
linens
Grapes, fresh or dried
Applications of Product Space Methodology in Turkish Economic Policy Making
The project attracted high amount of
media coverage and criticisms
What were the major shortcomings
of the project?
 City level export data is problematic
 Data is collected from customs and doesn’t
indicate the productive structure of cities
 Product space is created using country level export
data and embedded capabilities are relevant for
countries not for cities
 The results were communicated with the press as if
the Ministry has the magical formula for regional
development
Industrial policy strategy design for
Zafer Region
 We worked with Zafer Regional Development
Agency
 The purpose of the project was to identify
potential industries that could emerge with the
existing set of capabilities
 Informing the local competitiveness agenda
 Ambition: regional industrial policy strategy
 We implemented the product space techniques
but paid attention not to repeat the mistakes of
the previous project
 Developed new techniques and approaches along
the way!
Overview of the Zafer region
 4 provinces: Manisa, Kütahya, Afyon, Uşak
 2.8 million people (Albania), 4% of Turkey’s population
 USD 5 billion in exports (2012)
 24 companies in «Turkey 1000 industrial list»
 170 thousand non services employment
 Main industries
 Manisa: consumer electronics, home appliances
 Afyon: marbles
 Uşak: textile and leather
 Kütahya: ceramics and porcelain
 Limited backward and forward linkages
Manisa
OSB
Manisa
OSB
Kütahya
OSB
Akhisar
OSB
Uşak
OSB
Afyon
OSB
Product Space of the Zafer Region – 4 cities
Manisa
Kütahya
Afyon
Uşak
Challenge: A large variation in economic
diversity within the region
Share of top 10 sector in
employment, 2011
Rank in Turkey,
2011
MANİSA
%39.5
3
AFYON
%53.7
22
KÜTAHYA
%55.1
27
UŞAK
%66.0
49
Positioning of the 4 Zafer provinces in
vis-a-vis rest of Turkey
Positioning of all Zafer counties in the
policy matrix: ubiquity and open forest
Filtering to find the good products method
 Use of Employment Data, 2011
 Filter 1: Locality
 There should be some local capacity, but no stong
presence (LQ < 1)
 Filter 2: Proximity
 Jumping should be to the nearby products, with
density higher than average
 Filter 3: Quality
 Quality should be higher than the region’s average
quality
Caution: export vs. employment data
Export data
Advantages:
 Manufacturing industry is represented
by more sectors
 Global analysis is possible
 Easy to access.
Disadvantages:
 Provinces do not export everything they
produce
 Registration problem (customs)
Methodology:
 Compares economic structure of each
city to that of developed countries
 Quality sectors are defined as sectors
in advanced countries but not in the
devleloping ones.
 Draws a roadmap for provinces to catch
up with developed countries
Employment data
Advantages:
 A better representation of city’s
production structure
 Lower level of missing data arising
from registration problem
 Covers services sector
Disadvantages:
 Manufacturing industry lacks
sectoral representation.
 Not able to make global analysis
 Only for registered (formal)
employment
Methodology:
 The economic structure of provinces
are compare to each other.
 Quality sectors are defined as
sectors in advanced provinces but
not in the developing ones.
 Draws a roadmap for backward
provinces to catch up with developed
provinces such as İstanbul and
Bursa
Filtering to find the good products results
Localization
Filter
Proximity
Filter
Quality
Filter
Sektör adı
Sıradanlık
Yakınlık
Örme (Trikotaj) ve tığ işi diğer giyim eşyası imalatı
4
%14.1
Örgü ve tığ işi kumaşların imalatı
6
%15.3
Haşere ilaçları ve diğer zirai-kimyasal ürünlerin imalatı
7
%13.8
Oluklu kağıt ve oluklu mukavva imalatı ile kağıt ve mukavvadan yapılan
ambalaj kutuları imalatı
10
%12.8
Diğer baskı ve hakkaklık işleri (tabaklar ve diğer eşya üzerine baskı.
hakkaklık ve işleme yapılması gibi)
10
%13.6
Doğal ve sentetik pamuk elyafının hazırlanması ve eğrilmesi
10
%17.5
Dokuma olmayan kumaşlar ile dokuma olmayan kumaştan yapılan ürünlerin
imalatı. giyim eşyası hariç
11
%14.3
Plastik torba. çanta. poşet. çuval. kutu. damacana. şişe. makara vb.
paketleme malzemelerinin imalatı
12
%13.4
Kağıt ve mukavva imalatı
12
%14.9
Halat. ip. sicim ve ağ imalatı
12
%17.9
Başka yerde sınıflandırılmamış diğer özel amaçlı makinelerin imalatı
13
%12.6
Diğer kauçuk ürünleri imalatı
Metallerin dövülmesi. preslenmesi. baskılanması ve yuvarlanması, toz
metalürjisi
13
%13.0
13
%13.7
Ağaç mantarı ürünleri imalatı, saz. saman ve benzeri malzemelerden örülerek
yapılan ürünlerin imalatı
13
%14.4
Giyim eşyası dışındaki tamamlanmış tekstil ürünlerinin imalatı
13
%16.1
Ana demir ve çelik ürünleri ile demir alaşımları imalatı
14
%12.8
Sabun ve deterjan. temizlik ve parlatıcı maddeleri imalatı
14
%12.8
Nişasta ve nişastalı ürünlerin imalatı
14
%13.2
Metalden diğer tank. rezervuar ve büyük muhafaza kapları (kapasitesi
&gt,=300 litre) imalatı
15
%13.1
Düz cam imalatı
16
%13.3
Motosiklet imalatı
16
%14.4
Hafif metalden paketleme malzemeleri imalatı
16
%16.0
Source: Turkstat Tepav Calculations
Main findings from the Zafer Region Study
 New / potential products identified at the province
and county level
 e.g. Kütahya province – accumulator and battery
production (?)
 Caveat: Need deeper value chain analyses to see
whether this is really feasible
 Framework conditions & capabilities
 Investment Climate Assessment
 Survey conducted at the region level
 World Bank’s methodology, but modified to the
region’s context and our study’s objectives
 Key issues: high costs, access to finance, new product
development, human resources
Results from the Zafer Region’s Study –
Investment climate assessment
Tüm Firmalar
Mikro Firmalar
Maliyet Düşürme
Yatırım İçin Ek Finans
Yeni Ürün Geliştirme
Nitelikli Uzman Personel
İç Pazarda Büyüme
Kalitenin İyileştirilmesi
İlave İşletme Sermayesi
Teknoloji İyileştirme
Çalışanların İşbaşı Eğitimi
Markalaşma
Otomasyon
Yurtiçi Ortaklık
Dış Pazara Açılma
Yurtdışı Ortaklık
Yönetici Eğitimi
İşletmenin Yeniden Yapılanması
Elektronik Ticarete Geçiş
Yeni Bir Alana Girme
0
5
10
15
20
25
Main recommendations – inputs for the
local competitiveness agenda
 Value Chain Analysis Series (to complement product







space)
Export accelerator program (pilot countries,
matchmaking & networking)
Mobility grant program (to attract retired CEOs from
Istanbul)
Industrial technology centers (for key products)
A program to activate local universities to focus on
solutions to local problems
Celebrating local successes on innovation
Internationalization program for SMEs (training on
foreign languages)
Training for SMEs to benefit more from national grant
programs on R&D
Increasing the competitiveness of
Turkish machinery industry
 A project that we do for TOBB with the support of
the Ministry of Economy
 Local chambers are also involved (Institute of
Collaborations)
 Approximately 200 machinery producers from 13
different cities
 We are conducting needs assessments for each
and every company
 What do they need to become more competitive
 Needs assessment reports of TEPAV will form the
basis for Ministries’ cluster specific supports
Regional Industrial Policy Study
Globally machinery sector is at the heart of the
product space
Using Feenstra
et al. Trade Data:
132 Countries
1006 Products
SITC-4
(1975-2000)
CA Hidalgo, B Klinger, A-L Barabasi, R Hausmann Science (2007)
Regional Industrial Policy Study
Turkey lags behind key actors in the
machinery sector
Machinery Sector on Product Space for selected countries (2010)
The focus sectors were selected from among 91
machinery sub sectors
5 step selection process to determine
Machinery sub sectors
0
2
1
3
Importance
on global
trade
Importance
on Turkey’s
Export
Competition
Arena
(number of
sectors)
Large margin
growing sectors
in global trade
Large margin
growing sectors
in Turkey’s
trade
Sectors where
we can gain a
competitive
advantage
91
0
Source:Tepav Calculations
75
16
71
4
39
32
4
Import
Density
Access to
market
Sectors with
low barriers to
market entry
29
10
10
39
+ =
39 sectors survived that filtering process
Why is it important to gain competitiveness in
these 39 sub-sectors?
Source: Comtrade, Tepav Calculations
We’re making our hands dirty
What do we have in needs assessment
reports?
 External economic environment analysis
 Sector analysis (global trends)
 Firm structure and performance analysis
 Value chain assessment
 Road map for gov’t supports
 Consultancy services
 Education services
 Overseas marketing activities (fair
participation, customer visits etc…)
The ultimate goal is to help emulator and innovative
SMEs to become fast growing companies