Supplementary Material for : “Endogenous (In)Formal Institutions.”

Serra Boranbay (Carnegie Mellon University) and Carmine Guerriero (ACLE) August 3, 2016

Robustness to Alternative Assumptions A Natural Number of Agents Suppose that the number of agents is N , where N is odd and a majority M ≥ citizens. Let fe =

f M

N +1 2

of them are

be the upfront payment each citizen receives while investing. We are seeking

conditions parallel to those recorded in assumption 1 assuring that our analysis continues to go through. Starting from 1b and 1c, suppose that the elite picks investment and pe under autocracy. Given that k < M citizens have already cooperated, a citizen does not cooperate if u

k M γλ




−u

k−1 M γλ




< fe − d.

Since u00 < 0, the left hand side of this inequality is maximized at k = 1, i.e., there is a free rider effect such that, if no one else cooperates, the gains from cooperating are the highest. For cooperation not to occur, we set k = 1 in the previous inequality and assumption 1c now reads as  

−1 fe − d . Similarly, the new assumption 1b can be derived from u k λ −u k−1 λ < fe−d γ
fe−d . λ u M

Finally, assumption 1a rewrites as fe > d > λ > 1.



Relaxing Assumption 1 In the following, we relax one sub-assumption at a time, highlighting the effect in the process. 1a) Here, we consider each of the inequalities part of assumption 1a in turn:

— assuming f ≤ d entails that cooperating in the investment can be a dominant strategy for the citizenry. This helps democratization by relaxing both incentive compatibility constraints and, so, mitigating both sides’ needs of strategically manipulating culture. — first of all, maintaining d ≤ λ implies that cooperative risk-sharing may no longer be feasible if λR exceeds what a group can afford, i.e., d < λR . If so, then the elite’s payoff

2

from cheating in risk-sharing is zero, and thus s∗D (0, 0) is the only feasible tax rate. Hence, the choice of the political regime is no longer shaped by de and dc . Finally, d ≤ λ does not allow over-accumulation of culture and the related commitment dimension. — assuming λ ≤ 1 has a twofold effect on the equilibrium. On the one hand, it reduces the appeal of cheating one’s partner in risk-sharing and, as a result, relaxes (Ie ) and increases the prospects of democracy. On the other hand, it makes more likely that the citizenry expropriates f , and thus that neither λI or pc is produced.

1b); 1c) Under the weaker restrictions θc u λ < f − d∗c and u γλ < f − d∗c , there are dc > d∗c making the citizenry’s cooperation in investment feasible under autocracy but not chosen because either too costly or unable to satisfy constraint (Ie ). Hence, our analysis remains the same.

Risk-sharing Produces a Taxable Value Suppose that cooperation in risk-sharing creates a taxable value c, which can then be converted into public good production. This time, the group in charge fixes one tax rate for risk-sharing, denoted with j , and another for investment, labeled again sj . Since now public good production does not necessitate the upfront payment, it materializes under autocracy provided that u (A c) + (1 − A ) c+de > λR and θc u (A c)+(1 − A ) c+dc > λR (see tables I to IV). Hence, the elite prefers risk-sharing to investment and avoids granting democracy if λI and c are not too disparate.



Allowing Transfers Under Autocracy If the elite can direct residual funds to the citizenry under autocracy and thus elicit cooperation in this way, democratization is less likely. In addition to assumption 1 indeed, our results go through if also transfers are less valuable to the citizenry than either public good when supplied under autocracy because then they are unsolicited to begin with. In other words, a public good marginal

3



benefit to the citizenry under autocracy must be greater than one or min θc u0 λ , u0 γλ > 1. Assuming instead that citizenry can appropriate residual funds under democracy would collide with our maintained assumption that the untaxed investment value cannot be expropriated ex post. 

Different Timing of Events When culture does not impact the choice of the economic activity—i.e., in ranges (A) or (B), the timing of events is irrelevant. In the intermediate ranges of λI instead, deciding culture earlier is crucial since this choice establishes the elite’s payoff from investment and thus can permit credible commitment when needed. For example, if the elite selected the activity before culture was accumulated, then the citizenry could no longer commit through his choice of dc as it would not be subgame perfect. Anticipating this, the elite would not introduce democracy in the first place. Similarly, the tax rate should be decided before the elite selects the activity for our results to go through. Otherwise, the citizenry can no longer be trusted to pick a tax rate that makes constraint (Ie ) hold after investment since it would no longer be necessary for sequential rationality. Hence, the timing of events on which we focus in the basic setup is the one that resolves the credibility issues, encourages democratization, and is preferred by the elite who is also likely to select it. Any other order of events renders democratization impossible when λI falls in ranges (C), (D), and (E). To illustrate these points in the cleanest way, we consider here the case in which the elite selects the economic activity at time zero. Then, the new timing of events is the following: 0. the elite chooses the activity; 1. both groups accumulate their culture; 2. the elite selects the political regime; 3. the tax rate is set; 4. the agents are randomly matched, and possibly public good production follows investment. Assumptions 1 and 2 then immediately imply that the only viable alternatives are “autocracy and risk-sharing” and “democracy and investment.” In other words, the elite does not choose investment only to retain autocracy or risk-sharing to introduce democracy.

4

At time 3, the citizenry always fixes s∗D = 1 and cooperates in investment at time 4 only if   Ibc : u (λI ) ≥ f − d∗c . Since there is no risk-sharing outside option, the elite also cooperates in   investment if Ibe : θe u (λI ) ≥ f − d∗e . At time 2, the elite then chooses democracy if and only   if both inequalities hold. This is always the case for λI ≥ u−1 θfe , which is smaller than the threshold limiting range (B) but larger than that defining range (A) in the basic setup because now culture does not shape both parties’ investment opportunity cost and thus cannot permit credible commitment if needed. At time 1, when λR ≤ 1 and θe u (λI ) ≥ f − λR , which correspond to range (B) in the basic setup, d∗c = d∗e = 1 and both democracy and investment take place. For λR > 1 and θe u (λI ) ≥ f − λR , which correspond to range (C) in the basic setup, the potential candidates are (d∗c , d∗e ) ∈ {(λR , λR ) , (µ, λR ) , (λR , 1 − µ) , (0, 0)}, and the elite grants democracy given any such pair. Thus, the citizenry’s expected utility is λR + µu (λI ) −

λ2R 2

for d∗c = µ. Analogously, the elite’s expected utility is λR + (1 − µ) [θe u (λI ) − f ] − and (1 − µ) [θe u (λI ) − f + 1 − µ] − for λR ≤ (>) 1 +

(1−µ)2 2

2

for d∗c = λR and µ (u (λI ) + µ) − µ2 λ2R 2

for d∗e = λR

for d∗e = 1 − µ. Hence, the elite selects d∗e = λR (1 − µ)

p p 2µ − µ2 , and the citizenry sets d∗c = λR (µ) for λR ≤ (>) 1 + 1 − µ2 .

This means that a high λR has a more adverse effect on democracy than in the basic setup.   p Indeed, for λR > 1 + 2µ − µ2 and f − λR ≤ u (λI ) < f − (1 − µ), constraint Ibe does not hold. Similarly, for u (λI ) < f − (1 − µ) and f − λR ≤ θe u (λI ) < f − µ, which correspond to a subset   of range (D) in the basic setup, constraint Ibc fails and neither the commitment dimension of cultural accumulation nor democracy arise. The same conclusions are true for θe u (λI ) < f − λR , which corresponds to ranges (A) plus (E) in the basic setup. Finally, under autocracy culture maximizes the risk-sharing payoff only, and it equals the values prevailing in the basic setup.



Dynamics We assume that the basic setup is augmented by a period T in which emerges from the citizenry

5

a size β → 0 group of “ennobled merchants” able to produce without upfront payment an investment surplus βαλI,T > 0—where α > 0 and λa,T is the activity-specific factor a in the new period T — and to restore autocracy together with the elite and without bearing any cost. At the beginning of time T , democracy is in place, λI,T < λI , and the elite members (citizens and ennobled merchants) inherit a stock of culture (1 − δ) de ((1 − δ) dc ). For the sake of simplicity, we maintain that µ → 0, δ = 1, and before time T everybody believes that λa,T = λa for all a and ignores the rise of the ennobled merchants, which in turn can only engage in the new activity.

Under autocracy, the elite gives up the possibility to invest with the citizenry and levies on the ennobled merchants taxes 1−µ−β 1−µ βαλI,T

µ µ+β βαλI,T .

Under democracy instead, the citizenry collects taxes

from the ennobled merchants and tries to either share risk or invest with the elite.

Since the ennobled merchants are worst off when taxed by the citizens, they let the elite choose whether seizing power. A coup is always triggered if λI,T is in either ranges (A) or (E), and thus the elite’s payoff is weakly lower under investment, since then autocracy assures the elite the risk-sharing payoff plus the revenues from taxing the ennobled merchants. In this case, the elite and the citizenry build a culture maximizing only the risk-sharing payoff. If λI,T is in range (C) instead, the elite chooses de,T = 0 being µ → 0, and gets an utility equal to λR,T when the citizenry selects a culture that credibly signals his cooperation in investment. By comparing the elite’s utility from democracy—i.e., (1 − µ − β) λR,T —to the one from autocracy—i.e., leading to a coup is

λI,T λR,T

>

µ+β−(µ+β)2 . µβα

µ µ+β βαλI,T ,

the condition

If it holds, then the elite initiates a coup irrespective of the

citizenry’s choice of culture dc,T and each group builds a culture maximizing the risk-sharing payoff only. To illustrate, the elite’s choice is as in range (A) of the basic setup, whereas the citizens’ one differs only for the probability of meeting another citizen, which is now 1 − µ − β. If the condition for a coup fails, the citizenry has a very good reason to sustain democracy since otherwise he would lose both

1−µ−β 1−µ βαλI,T

and f . If affordable, he builds a dc,T such that

6

(1 − µ) dc,T Idc,T ≥λR,T + µf +

where πR,c,e is

(1−µ−β)2 2

1−µ−β 1−µ βαλI,T

−

d2c,T 2

if λR,T ≤ 1 − µ − β, (1 − µ − β) λR,T −

≥ πR,c,e ,

(λR,T ) 2

2

if 1 − µ − β < λR,T ≤

2 (1 − µ − β), and 0 otherwise. The case in which λI,T belongs to range (B) is similar. Being investment more lucrative, there are no commitment issues by the citizenry, and thus reviving autocracy is harder. Indeed, the elite triggers the coup only if (1 − µ − β) [θe u (λI,T ) − f + de,T ] < µ µ+β βαλI,T .

If this condition is not true, democracy reigns. Finally, if λI,T belongs to range (D), it

is significantly harder for democracy to persist because of the small investment payoff.



Data on Political Institutions The POLITY IV dataset codes various “institutionalized authority” characteristics of political regimes across the world after the year 1800 (Marshall and Jaggers, 2016). Our measure of the inclusiveness of political institutions corresponds to the variable “constraint on the executive,” which ranges between 1 and 7 with higher values indicating that the holder of executive powers is accountable to the citizenry and/or the government is constrained by checks and balances. To elaborate, a score of 1 points at a situation in which there are no regular limitations on the executive’s actions distinct from irregular limitations such as the threat or actuality of coups. “Absolutist monarchies, regardless of their openness to public dissent or respect for civil liberties, are typically coded here” (Marshall and Jaggers, 2016). A value of 3 describes executives that face real but limited constraints. For instance, there is a legislative body that has more than just consultative functions, but can also delay implementation of executive decrees, or can initiate some categories of legislation. A score of 5 corresponds to an executive having more effective authority than any accountability group but subject to substantial constraints by them. Examples are a legislature that often modifies or defeats executive proposals for action; a council or legislature that sometimes

7

refuses funds to the executive; an accountability group that makes important appointments to administrative posts. A value of 7 captures a situation in which accountability groups have effective authority equal to or greater than the executive in most public activities. Finally, the intermediate scores of 2, 4 and 6 correspond to transitions between the above situations. To assign a “constraint on the executive” score to each observation in our sample, we proceed as follows. If a state did not show sizable heterogeneity in the inclusiveness of political institutions between 1000 and 1600, we attribute to all its historical regions the score devised by Acemoglu et al. (2005) if sufficiently accurate. This choice takes care of Austria, Ireland, Portugal, and the UK. In the cases of sizable within-state heterogeneity in the inclusiveness of political institutions—i.e., Belgium, France, Germany, Italy, Poland, and Spain, inaccuracy of the Acemoglu et al.’s (2005) codification—i.e., Czech Republic, Hungary, Netherlands, and Switzerland—and of states not covered by these Authors—i.e., Slovenia and Slovakia, we use the POLITY IV definition to extrapolate from the historical facts in a 40 year window around each date a score for each observation (see also Acemoglu et al., [2005]). Our main sources are Galasso (1976), Stearns (2001), and Ortu (2005).

References Acemoglu, Daron, Simon Johnson, and James A. Robinson. 2005. “The Rise of Europe: Atlantic Trade, Institutional Change and Economic Growth.” American Economic Review, 95: 546-579.

Conley, Tim G. 1999. “GMM Estimation With Cross Sectional Dependence.” Journal of Econometrics, 92: 1-45.

Derbyshire, James, Ben Gardiner, and Sevrin Waights. 2013. “Estimating the Capital Stock for the NUTS2 Regions of the EU27.” Applied Economics, 45: 1133-1149.

8

Galasso, Giuseppe. 1976. Storia d’Italia. Torino: UTET.

GESIS. 2008. EVS 2008 - Variable Report Integrated Dataset. K¨oln, Germany: GESIS Press.

Guiot, Joel, Christophe Corona, and ESCARSEL members. 2010. “Growing Season Temperatures in Europe and Climate Forcings Over the Past 1400 Years.” PLoS ONE, 5: e9972.

Luterbacher, J¨ urg, Daniel Dietrich, Elena Xoplaki, Martin Grosjean, and Heinz Wanner. 2004. “European Seasonal and Annual Temperature Variability, Trends, and Extremes Since 1500.” Science, 303: 1499-1503.

Marshall, Monty G., and Keith Jaggers. 2016. “Polity IV Project, Political Regime Characteristics and Transitions 1800-1999, Dataset Users Manual.” Center for International Development and Conflict Management, University of Maryland.

Moorman, John R.H. 1983. Medieval Franciscan Houses. New York, NY: Franciscan Institute Publications.

Ortu, Gian Giacomo. 2005. La Sardegna Dei Giudici. Nuoro: Il Maestrale.

Stearns, Peter N. 2001. The Encyclopedia of World History: Ancient, Medieval, and Modern, Chronologically Arranged. Cambridge: James Clarke and Co.

Tabellini, Guido. 2010. “Culture and Institutions: Economic Development in the Regions of Europe.” Journal of European Economic Association, 8: 677-716.

Van Der Meer, Frederic. 1965. Atlas de L’Ordre Cistercien. Amsterdam: Elsevier.

9

Tables Table I: The Modified Risk-sharing Game Under Autocracy When pe is Chosen Citizen

Cooperate Non Cooperate

Elite Cooperate θc u (A c) + dc , u (A c) + (1 − A ) c + de λR , d e − λR

Non Cooperate dc − λR , λR 0, 0

Table II: The Modified Risk-sharing Game Under Autocracy When pc is Chosen Citizen

Cooperate Non Cooperate

Elite Cooperate u (A γc) + dc , θe u (A γc) + (1 − A ) c + de λR , d e − λR

Non Cooperate dc − λR , λR 0, 0

Table III: The Modified Risk-sharing Game Under Democracy When pc is Chosen Citizen

Cooperate Non Cooperate

Elite Cooperate u (A c) + dc , θe u (A c) + (1 − A ) c + de λR , d e − λR

Non Cooperate dc − λR , λR 0, 0

Table IV: The Modified Risk-sharing Game Under Democracy When pe is Chosen Citizen

Cooperate Non Cooperate

Elite Cooperate θc u (A γc) + dc , u (A γc) + (1 − A ) c + de λR , d e − λR

Non Cooperate dc − λR , λR 0, 0

Table V: The Sample — Matching Between NUTS 2 and Historical Regions Styria-Austria (Burgenland, Nieder¨ osterreich, Ober¨ osterreich; Steiermark; Wien); Tyrol - Trentino-Alto Adige (K¨ arnten, Salzburg, Tirol, Vorarlberg, Bolzano, Trento); R´ egion Bruxelles (R´ egion Bruxelles, Vlaams-Brabant); R´ egion Wallone (Brabant Wallon, Hainaut, Li` ege, Luxembourg, Namur); Vlaams Gewest (Antwerpen, Limburg, Oost-Vlaanderen, West-Vlaanderen); East Czech Republic (Jihovychod, Moravskoslezsko, ˇ Stredni Morava); West Czech Republic (Stˇ redn´ı Cechy, Severozapad, Severovychod, Jihozapad, Praha); Corse; East France (Alsace, FrancheComt´ e, Lorraine, Rhˆ one-Alpes); ˆ Ile de France; Mediterranean France (Languedoc-Roussilon, Provence-Alpes-Core D’azur); North France (Champagne-Ardenne, Picardie, Haute-Normandie, Basse-Normandie, Nord-pas-de-Calais, Pays de la Loire, Bretagne); Paris Basin (Limousin, Centre, Bourgogne); South-East France (Auvergne); South-West France (Midi-Pyr´ en´ ees); West France (Poitou-Charentes, Aquitaine); BadenW¨ urttemberg (Stuttgart, Karlsruhe, Freiburg, T¨ ubingen); Bayern (Oberbayern, Niederbayern, Oberpfalz, Oberfranken, Mittelfranken, Unterfranken, Schwaben); Brandenburg (Berlin, Brandenburg Nordost, Brandenburg S¨ udwest); Bremen - Hamburg - Niedersachsen (Bremen, Hamburg, Braunschweig, Hannover, L¨ uneburg, Weser-Ems); Hessen (Darmstadt, Gießen, Kassel); Mecklenburg-Vorpommern; Nordrhein-Westfalen (D¨ usseldorf, K¨ oln, M¨ unster, Detmold, Arnsberg); Rheinland-Pfalz - Saarland (Koblenz, Trier, Rheinhessen-Pfalz, Saarland); Sachsen (Chemnitz, ´ Dresden, Leipzig); Schleswig-Holstein; Th¨ uringen - Sachsen-Anhalt; Central Hungary (K¨ oz´ ep-Magyarorsz´ ag, Eszak-Magyarorsz´ ag); Styria-Hungary ´ (Eszak-Alf¨ old, D´ el-Alf¨ ol); West Hungary (K¨ oz´ ep-Dun´ ant´ ul, D´ el-Dun´ ant´ ul, Nyugat-Dun´ ant´ ul); East Ireland (Southern and Eastern); West Ireland (Border, Midland and Western); Abruzzo - Molise; Basilicata - Campania; Calabria; Emilia-Romagna; Friuli-Venezia Giulia - Veneto; Lazio; Liguria; Lombardia; Marche - Umbria; Piemonte - Valle D’Aosta; Puglia; Sardegna; Sicilia; Toscana; Noord Nederland - Groningen (Drenthe, Flevoland, Friesland, Gelderland, Groningen, Noord-Holland, Overijssel); Oost-Nederland (Limburg); West-Nederland (Zeeland); Zuid-Nederland (Noord-Brabant, Utrecht, Zuid-Holland); East Poland (Lubelskie, Mazowieckie, Podlaskie, Warminsko-Mazurskie); North Poland (L´ odzkie, Kujawsko- Pomorskie, Pomorskie, Wielkopolskie); South Poland (Malopolskie, Dolnoslaskie, Opolskie, Podkarpackie, Slaskie, Swietokrzyskie); West Poland (Lubuskie, Zachodniopomorskie); Alentejo; Algarve; Centro; Lisboa - Vale do Tejo; Norte; East Slovakia (Bratislavsk´ y Kraj); West Slovakia (V´ ychodn´ e Slovensko, Z´ apadn´ e Slovensko, Stredn´ e Slovensko); Carniola (Vzhodna); Styria-Slovenia (Zahodna); Andalucia; Aragon; Asturias - Cantabria; Baleares; Castilla-La Mancha; Castilla y Le´ on; Catalu˜ na; Comunidad Valencian; Extremadura; Galicia; Madrid; Murcia; Navarra - Rioja; Pais Vasco; North Switzerland (Nordwestschweiz, Z¨ urich, Ostschweiz); South Switzerland (Espace Mittelland, R´ egion L´ emanique, Ticino, Zentralschweiz); East Anglia - London (London-Inner, London-Outer, East Anglia, Bedfordshire and Hertfordshire, Essex); East Midlands (Derbyshire and Nottinghamshire, Leicestershire, Rutland and Northamptonshire, Lincolnshire); North-East UK (Tees Valley and Durham, Northumberland and Tyne and Wear); North-West UK (Cumbria, Cheshire, Greater Manchester, Lancashire, Merseyside); Northern Ireland; Scotland (Eastern Scotland, South Western Scotland, North Eastern Scotland, Highlands and Islands); South-East UK (Surrey, East and West Sussex, Hampshire and Isle of Wight, Kent, Berkshire, Buckinghamshire and Oxfordshire); South-West UK (Dorset and Somerset, Cornwall and Isles of Scilly, Devon, Gloucestershire, Wiltshire and Bristol/Bath area); Wales (Wales-West, Wales-East); West Midlands (Herefordshire, Worcestershire and Warwickshire, Shropshire and Staffordshire, West Midlands); Yorkshire and the Humber (East Yorkshire and Northern Lincolnshire, North Yorkshire, South Yorkshire, West Yorkshire). Note: 1. The names of the present-day NUTS 2 regions are in regular font, and those of the historical regions to which they belonged are in Italic type.

10

Table VI: Summary of Variables Variable Houses-1000-1600 : Past Culture:

Houses-C-1000-1600 : Houses-F-1000-1600 : Temp-SD:

Geography:

Temp-A: East:

See text.

West: Alternative mechanisms:

Note:

1.

Definition and Sources Number of Cistercian and Franciscan houses per square km in the region. Sources: http://www.cistercensi.info/; Van Der Meer (1965); http://users.bart.nl/∼roestb/franciscan; Moorman (1983). Number of Cistercian houses per square km in the region. Sources: http://www.cistercensi.info/; Van Der Meer (1965). Number of Franciscan houses per square km in the region. Sources: http://users.bart.nl/∼roestb/franciscan; Moorman (1983). Standard deviation of the growing season temperature over the previous 50 years in Celsius averaged over the cells in the region. Source: Guiot et al. (2010) for the 950-1500 period and Luterbacher (2004) for the remainder of the sample. Mean growing season temperature over the previous 50 years in Celsius anomalies relative to the 1961-1990 mean averaged over the cells in the region. Source: Guiot et al. (2010) for the 950-1500 period and Luterbacher (2004) for the remainder of the sample. See text.

Statistics 0.0018 (0.0022) (0.0022) 0.0003 (0.0002) 0.0015 (0.0021) 1.068 (1.489) 2.110 (4.888) 257.230 (509.297) 1827.438 (2464.532) 0.268 (0.135)

Share of respondents declaring themselves Roman Catholic and answering “very important” to the question “how important is religion in your life?” averaged over the cells in the region. Source: 2008 European Value Study, GESIS (2008). 2000 real capital stock per capita in millions of euro averaged over the cells in the 0.054 Capital: region. Source: Derbyshire et al. (2013). (0.024) The last column reports the mean and, in parentheses, the standard deviation of each variable. Both are computed employing the sample used in tables VII and XII except for the case of Temp-SD and Temp-A, when they are calculated using the sample on which table IX is based. Catholicism:

Table VII: Persistent Culture — Measuring Past Culture With the Numbers of Monasteries (1)

(2)

(3)

(4)

(5) (6) The dependent variable is

Culture-2008 Houses-1000-1600 Houses-C-1000-1600

13.193 (5.161)**

(7)

Hardwork-2008 3.258 (2.357)

220.651 (135.769)*

(8) Thrift-2008

- 2.513 (6.410) 60.059 (64.936)

- 21.922 (71.250)

10.281 2.459 (5.985)* (2.973) Fixed country effects YES YES YES YES YES YES YES YES Estimation OLS. R2 0.16 0.18 0.16 0.06 0.07 0.06 0.14 0.14 Number of observations 89 89 89 89 89 89 89 89 Notes: 1. Standard errors clustered at the country level in parentheses. *** denotes significant at the 1% confidence level; **, 2. The specifications always include also Coast, Area, Latitude, and Longitude. Houses-F-1000-1600

11

(9)

- 2.249 (6.485) YES 0.13 89 5%; *, 10%.

Table VIII: Persistent Culture — Individuals Samples (1)

(2)

(3)

(4) (5) (6) (7) (8) Panel A: Full sample. The dependent variable is

Culture-2008 Culture-1000-1600

1.518 (0.540)**

- 0.020 (0.024)

0.106 (0.162) - 0.030 (0.029)

YES

YES

YES

- 0.016 (0.027) 0.430 (0.260)* YES

0.03 19624

0.03 19624

0.03 19624

0.03 19624

(1)

(2)

(3)

Temperature-SD-1000-1600

0.015 (0.010)

YES OLS. 0.02 0.01 20210 20210

0.009 (0.009) - 0.100 (0.079) YES

0.033 (0.021)

0.009 (0.137) 0.034 (0.021)

YES

YES

0.033 (0.020) - 0.049 (0.086) YES

0.02 20210

0.02 20049

0.02 20049

0.02 20049

(9)

(10)

(4) (5) (6) (7) (8) Panel B: Catholics Only. The dependent variable is Hardwork-2008

1.648 (0.612)**

Culture-F-1000-1600

Temperature-SD-1000-1600

0.038 (0.052) 0.012 (0.009)

Thrift-2008

0.124 (0.135)

Culture-C-1000-1600

Coast

- 0.115 (0.558)

YES

Culture-2008 Culture-1000-1600

Thrift-2008

0.439 (0.415)

- 0.028 (0.028)

Culture-F-1000-1600

Fixed country effects Estimation R2 Number of observations

(10)

0.228 (0.121)*

Culture-C-1000-1600

Coast

Hardwork-2008

(9)

- 0.023 (0.032)

- 0.019 (0.031)

0.728 (0.443) 0.039 (0.193) - 0.024 (0.033)

- 0.005 (0.032) 0.642 (0.303)** YES

- 0.005 (0.014)

0.140 (0.504) 0.029 (0.039) - 0.007 (0.014)

- 0.009 (0.014) - 0.073 (0.079) YES

- 0.066 (0.144) 0.036 0.036 0.033 (0.012)*** (0.012)*** (0.011)*** - 0.081 (0.072) YES YES YES

Fixed country effects YES YES YES YES YES Estimation OLS. 2 R 0.10 0.10 0.10 0.10 0.28 0.28 0.28 0.03 0.03 0.03 Number of observations 10129 10129 10129 10129 10452 10452 10452 10326 10326 10326 Notes: 1. Standard errors clustered at the country level in parentheses. *** denotes significant at the 1% confidence level; **, 5%; *, 10%. 2. The specifications always include also Area, Latitude, Longitude, individual-level controls for age and age squared, individual-level dummies for males and married, and an individual-level educational attainment indicator. 3. Panel B considers only respondents reporting that they are Catholic.

12

Table IX: Endogenous Institutions — Dynamics and Alternative Controls Culture

(1)

(2)

(3)

(4)

Democracy - 0.276 (0.241)

Culture

Democracy

Culture

- 0.029 (0.024) 1.167 (0.499)** 1.072 (0.499)** 1.340 (0.473)*** 1.060 (0.410)*** 0.331 (0.317) 0.029 (0.347) - 0.219 (0.333) - 0.206 (0.333) - 0.215 (0.342) 0.452 (0.377) 0.400 (0.390) 0.003 (0.039) - 0.346 (0.209)* - 0.311 (0.230) - 0.387 (0.233)* 0.296 (0.220 0.143 (0.206) 0.188 (0.204) 0.386 (0.168)** 0.224 (0.195) - 0.123 (0.193) - 0.040 (0.271) 0.517 (0.336)

- 0.005 (0.004) - 0.010 (0.007) - 0.006 (0.004) - 0.003 (0.003) 0.009 (0.007) 0.035 (0.018)** 0.078 (0.038)** 0.159 (0.065)** 0.236 (0.091)*** 0.319 (0.118)*** 0.437 (0.160)*** 0.567 (0.193)*** 0.0005 (0.0014) - 0.006 (0.005) 0.003 (0.003) - 0.001 (0.003) - 0.003 (0.006) - 0.0001 (0.008) - 0.012 (0.015) - 0.019 (0.020) - 0.038 (0.029) - 0.047 (0.040) - 0.084 (0.077) - 0.115 (0.085)

Democracy Mediterranean × 1050 Mediterranean × 1100 Mediterranean × 1150 Mediterranean × 1200 Mediterranean × 1250 Mediterranean × 1300 Mediterranean × 1350 Mediterranean × 1400 Mediterranean × 1450 Mediterranean × 1500 Mediterranean × 1550 Mediterranean × 1600 Atlantic × 1050 Atlantic × 1100 Atlantic × 1150 Atlantic × 1200 Atlantic × 1250 Atlantic × 1300 Atlantic × 1350 Atlantic × 1400 Atlantic × 1450 Atlantic × 1500 Atlantic × 1550 Atlantic × 1600 Temperature-SD

- 0.112 (0.046)** 1.090 (0.503)** 0.901 (0.514)* 1.227 (0.486)*** 1.111 (0.411)*** 0.338 (0.318) 0.070 (0.361) - 0.144 (0.363) - 0.090 (0.373) - 0.154 (0.399) 0.315 (0.466) 0.219 (0.493) 0.101 (0.058)* - 0.216 (0.197) - 0.156 (0.210) - 0.303 (0.226) 0.220 (0.228) - 0.011 (0.206) - 0.058 (0.210) 0.203 (0.173) 0.074 (0.194) - 0.325 (0.208) - 0.167 (0.238) 0.381 (0.307) 1.211 (0.424)***

Temperature-SD 2

- 0.012 (0.013) 0.011 (0.005)** 0.033 (0.018)* 0.032 (0.017)* 0.024 (0.019) 0.012 (0.019) 0.035 (0.021)* 0.086 (0.036)** 0.130 (0.061)** 0.208 (0.087)** 0.331 (0.116)*** 0.477 (0.153)*** 0.608 (0.190)*** - 0.007 (0.005) - 0.008 (0.008) - 0.019 (0.009)** - 0.008 (0.008) 0.017 (0.010)* 0.011 (0.014) 0.014 (0.019) - 0.009 (0.025) - 0.020 (0.031) - 0.043 (0.043) - 0.052 (0.055) - 0.091 (0.074) 0.563 (0.301)* - 0.359 (0.214)*

- 0.259 (0.201)

Temp-SD Temp-SD 2 ... Estimation Within R2 Observation number

Notes:

1. 2.

(. . . )

(. . . )

(. . . )

0.34 1170

0.58 1170

0.33 1170

(5) (6) The dependent variable is Culture-C Culture-F

0.0003 (0.0007) 0.0007 (0.0013) - 0.0001 (0.0011) - 0.002 (0.002) - 0.004 (0.005) - 0.003 (0.007) - 0.003 (0.010) - 0.005 (0.012) - 0.008 (0.014) - 0.010 (0.016) - 0.0001 (0.0184) 0.002 (0.020) - 0.00001 (0.00023) 0.0004 (0.0007) - 0.0003 (0.0008) 0.001 (0.002) 0.004 (0.004) 0.006 (0.007) 0.009 (0.009) 0.011 (0.011) 0.014 (0.013) 0.016 (0.016) - 0.015 (0.019) - 0.024 (0.020)

(7)

(8)

(9)

(10)

Democracy

Culture

Culture-C

Culture-F

- 0.129 (0.465)

0.373 (0.293) - 0.227 (0.211)

0.194 (0.090)** - 0.122 (0.080)

0.179 (0.295) - 0.105 (0.208)

(. . . )

(. . . )

(. . . )

0.54 1170

0.60 1170

0.47 1170

- 0.005 (0.004) - 0.010 (0.007) - 0.006 (0.004) - 0.001 (0.002) 0.012 (0.004)*** 0.039 (0.014)*** 0.081 (0.033)** 0.164 (0.060)*** 0.243 (0.086)*** 0.329 (0.111)*** 0.437 (0.153)*** 0.565 (0.187)*** 0.0005 (0.0014) - 0.006 (0.005) 0.004 (0.003) - 0.002 (0.002) - 0.007 (0.004)* - 0.006 (0.005) - 0.021 (0.010)** - 0.030 (0.015)** - 0.052 (0.024)** - 0.063 (0.035)* - 0.069 (0.075) - 0.092 (0.083)

- 0.224 (0.182) 0.022 (0.025) (. . . ) Fixed

0.021 - 0.245 (0.029) (0.181) - 0.007 0.029 (0.004)** (0.025) (. . . ) (. . . ) (. . . ) region and time effects OLS.

0.58 1170

0.61 1170

0.52 1170

0.33 1170

Standard errors clustered at the regional level in parentheses. *** denotes significant at the 1% confidence level; **, 5%; *, 10%. The specifications always include Temperature-A except for those in columns (3) to (6) that consider instead Temp-A.

13

Table X: Endogenous Institutions — Dynamics and Alternative Controls (Continued) ... East × 1050

(1)

(2)

(3)

(4)

Democracy (. . . )

Culture (. . . )

Democracy (. . . )

Culture (. . . )

(5) (6) The dependent variable is Culture-C Culture-F (. . . ) (. . . )

East × 1100

0.0007 (0.0004)* 0.0011 (0.0004)*** 0.0010 (0.0003)*** 0.0004 (0.0002)* 0.0002 (0.0003) -8E−06

East × 1150 East × 1200 East × 1250

East × 1350 East × 1400 East × 1450 East × 1500 East × 1550 East × 1600 West × 1050 West × 1100 West × 1150 West × 1200 West × 1250 West × 1300 West × 1350 West × 1400 West × 1450 West × 1500 West × 1550 West × 1600 Estimation Within R2 Observation number

1. 2.

(8)

0.00001 (7E−06 )* 5E−06 (5E−06 ) - 1E−06 (6E−06 )

0.00002 (0.00001) 0.00005 (0.00003)* 0.00008 (0.0003) (0.00005)* 0.00003 0.00013 (0.00026) (0.00007)* - 0.0001 0.0002 (0.0003) (0.0001)** 0.0002 0.0003 (0.0003) (0.0001)** 0.0001 0.0004 (0.0003) (0.0002)** 0.00002 - 2E−06 (0.00001)** (1E−06 ) - 0.00004 - 1E−06 (0.00004) (1E−06 ) - 0.00003 - 3E−06 (0.00004) (2E−06 )* - 0.00005 - 1E−06 (0.00005) (1E−06 ) 0.00005 2E−06 (0.00005) (2E−06 ) 0.00001 8E−07 (0.00004) (2E−06 ) 4E−06 3E−07 (0.00004) (4E−06 ) 0.00005 - 5E−06 (0.00004) (5E−06 ) 0.00003 - 8E−06 (0.00004) (6E−06 ) - 0.00004 - 0.00001 (0.00004) (9E−06 )* - 0.00002 - 0.00002 (0.00005) (0.00001)* 0.0001 - 0.00003 (0.0001) (0.00001)** Fixed region and time effects OLS.

East × 1300

Notes:

(7)

Democracy Culture (. . . ) (. . . ) - 0.00009 7E−06 (0.00004)*** (4E−06 )* 0.0009 10E−06 (0.0004)** (5E−06 )**

0.34 1170

0.58 1170

0.33 1170

0.58 1170

0.61 1170

0.52 1170

0.33 1170

0.54 1170

(9)

(10)

Culture-C (. . . ) 1E−06 (1E−06 ) 3E−06 (1E−06 )** 1E−06 (2E−06 ) - 2E−06 (3E−06 ) - 5E−06 (4E−06 ) - 6E−06 (6E−06 ) - 5E−06 (7E−06 )

Culture-F (. . . ) 6E−06 (3E−06 )* 7E−06 (5E−06 )

- 0.00001 (9E−06 ) - 0.00001 (0.00001) - 0.00001 (0.00001) - 9E−06 (0.00001) - 0.00001 (0.00001) - 5E−08 (4E−07 ) 3E−07 (5E−07 ) - 3E−07 (6E−07 ) 3E−07 (5E−07 ) 1E−06 (9E−07 ) 1E−06 (2E−06 ) 2E−06 (2E−06 ) 2E−06 (2E−06 ) 3E−06 (3E−06 ) 3E−06 (3E−06 ) 3E−06 (4E−06 ) 6E−08 (4E−06 ) 0.60 1170

0.00001 (6E−06 )* 6E−06 (4E−06 )* 3E−06 (3E−06 ) 0.00002 (0.00001)** 0.0001 (0.00002)** 0.0001 (0.00004)** 0.0001 (0.0001)** 0.0002 (0.0001)** 0.0003 (0.0001)*** 0.0004 (0.0001)*** - 2E−06 (9E−07 )* - 1E−06 (1E−06 ) - 3E−06 (1E−06 )* - 1E−06 (9E−07 ) 1E−06 (10E−07 ) - 4E−07 (2E−06 ) - 2E−06 (3E−06 ) - 7E−06 (4E−06 )* - 0.00001 (5E−06 )** - 0.00002 (8E−06 )** - 0.00002 (0.00001)** - 0.00003 (0.00001)** 0.47 1170

Standard errors clustered at the regional level in parentheses. *** denotes significant at the 1% confidence level; **, 5%; *, 10%. The specifications always include Temperature-A except for those in columns (3) to (6) that consider instead Temp-A.

14

Table XI: Endogenous Institutions — Allowing for Spatial Correlation (1)

(2)

(3)

(4) (5) (6) (7) The dependent variable is Democracy Culture Culture-C Culture-F Democracy Culture Culture-C - 0.115 0.012 0.002 0.011 - 0.115 0.012 0.002 Mediterranean × 1050 (0.050)** (0.008) (0.001) (0.007) (0.288) (0.059) (0.011) 1.086 0.020 0.004 0.016 1.086 0.020 0.004 Mediterranean × 1100 (0.063)*** (0.014) (0.003) (0.011) (0.289)*** (0.059) (0.011) 0.896 0.021 0.002 0.019 0.896 0.021 0.002 Mediterranean × 1150 (0.089)*** (0.015) (0.003) (0.013) (0.290)*** (0.059) (0.011) 1.226 0.009 - 0.002 0.011 1.226 0.009 - 0.002 Mediterranean × 1200 (0.049)*** (0.009) (0.002) (0.008) (0.289)*** (0.059) (0.011) 1.110 - 0.001 - 0.005 0.004 1.110 - 0.001 - 0.005 Mediterranean × 1250 (0.027)*** (0.006) (0.001)*** (0.005) (0.288)*** (0.059) (0.011) 0.329 0.031 - 0.005 0.036 0.329 0.031 - 0.005 Mediterranean × 1300 (0.001)*** (0.005)*** (0.001)*** (0.004)*** (0.288) (0.059) (0.011) 0.047 0.086 - 0.002 0.088 0.047 0.086 - 0.002 Mediterranean × 1350 (0.009)*** (0.002)*** (0.001)*** (0.002)*** (0.288) (0.059) (0.011) - 0.181 0.132 - 0.009 0.141 - 0.181 0.132 - 0.009 Mediterranean × 1400 (0.049)*** (0.010)*** (0.002)*** (0.008)*** (0.289) (0.059)** (0.011) - 0.148 0.210 - 0.010 0.219 - 0.148 0.210 - 0.010 Mediterranean × 1450 (0.063)** (0.007)*** (0.002)*** (0.006)*** (0.290) (0.059)*** (0.011) - 0.245 0.333 - 0.008 0.341 - 0.245 0.333 - 0.008 Mediterranean × 1500 (0.027)*** (0.006)*** (0.001)*** (0.005)*** (0.288) (0.059)*** (0.011) 0.186 0.475 - 0.005 0.480 0.186 0.475 - 0.005 Mediterranean × 1550 (0.072)** (0.016)*** (0.004) (0.013)*** (0.292) (0.060)*** (0.011) 0.053 0.607 - 0.006 0.613 0.053 0.607 - 0.006 Mediterranean × 1600 (0.086) (0.018)*** (0.004) (0.014)*** (0.291) (0.060)*** (0.011) 0.103 - 0.009 - 0.00002 - 0.009 0.103 - 0.009 - 0.00002 Atlantic × 1050 (0.044)** (0.007) (0.00100) (0.006) (0.251) (0.051) (0.00929) - 0.215 - 0.005 0.0019 - 0.007 - 0.215 - 0.005 0.0019 Atlantic × 1100 (0.041)*** (0.008) (0.0011) (0.008) (0.252) (0.051) (0.0093) - 0.152 - 0.017 - 0.0012 - 0.016 - 0.152 - 0.017 - 0.0012 Atlantic × 1150 (0.076)* (0.012) (0.0020) (0.011) (0.252) (0.052) (0.0093) - 0.302 - 0.005 0.002 - 0.006 - 0.302 - 0.005 0.002 Atlantic × 1200 (0.031)*** (0.006) (0.001)* (0.006) (0.251) (0.051) (0.009) 0.217 0.015 0.006 0.009 0.217 0.015 0.006 Atlantic × 1250 (0.045)*** (0.009) (0.002)*** (0.008) (0.251) (0.051) (0.009) - 0.012 0.011 0.006 0.004 - 0.012 0.011 0.006 Atlantic × 1300 (0.058) (0.013) (0.002)*** (0.012) (0.252) (0.052) (0.009) - 0.061 0.015 0.010 0.004 - 0.061 0.015 0.010 Atlantic × 1350 (0.110) (0.019) (0.003)*** (0.017) (0.255) (0.052) (0.009) 0.206 - 0.011 0.010 - 0.021 0.206 - 0.011 0.010 Atlantic × 1400 (0.070)*** (0.009) (0.001)**** (0.009)** (0.252) (0.052) (0.009) 0.079 - 0.021 0.014 - 0.035 0.079 - 0.021 0.014 Atlantic × 1450 (0.073) (0.008)** (0.001)*** (0.007)*** (0.252) (0.051) (0.009) - 0.314 - 0.040 0.015 - 0.055 - 0.314 - 0.040 0.015 Atlantic × 1500 (0.068)*** (0.012)*** (0.002)*** (0.012)*** (0.253) (0.052) (0.009)* - 0.152 - 0.050 0.016 - 0.067 - 0.152 - 0.050 0.016 Atlantic × 1550 (0.122) (0.018)** (0.003)*** (0.017)*** (0.256) (0.052) (0.009)* 0.407 - 0.096 0.002 - 0.098 0.407 - 0.096 0.002 Atlantic × 1600 (0.088)*** (0.018)*** (0.003)*** (0.018)*** (0.261) (0.053)* (0.010) - 0.024 0.550 0.199 0.351 - 0.024 0.550 0.199 Temperature-SD (0.572) (0.498) (0.113)* (0.393) (0.499) (0.338) (0.061)*** - 0.347 - 0.125 - 0.223 - 0.347 - 0.125 Temperature-SD 2 (0.397) (0.085) (0.317) (0.283) (0.051)** Estimation Fixed region and time effects OLS. Within R2 0.34 0.58 0.60 0.53 0.34 0.58 0.60 Number of observations 1170 1170 1170 1170 1170 1170 1170 Notes: 1. Driscoll-Kraay (Conleys (1999)) standard errors in the parentheses of columns (1) to (4) ((5) to (8)). *** denotes the 1% confidence level; **, 5%; *, 10%. 2. The specifications always include Temperature-A.

(8) Culture-F 0.011 (0.057) 0.016 (0.057) 0.019 (0.057) 0.011 (0.057) 0.004 (0.057) 0.036 (0.057) 0.088 (0.057) 0.141 (0.057)** 0.219 (0.057)*** 0.341 (0.057)*** 0.480 (0.058)*** 0.613 (0.058)*** - 0.009 (0.049) - 0.007 (0.050) - 0.016 (0.050) - 0.006 (0.049) 0.009 (0.049) 0.004 (0.050) 0.004 (0.050) - 0.021 (0.050) - 0.035 (0.050) - 0.055 (0.050) - 0.067 (0.051) - 0.098 (0.051)* 0.351 (0.326) - 0.223 (0.273) 0.53 1170 significant at

Table XII: Persistent Culture — Alternative Mechanisms (1) Culture-1000-1600

(2)

0.217 (0.080)** 1.732 (1.084)

Culture-C-1000-1600

- 0.131 (0.063)*

0.128 (0.086) - 0.140 (0.066)**

1.041 (1.208)

0.989 (1.157)

1.150 (1.259)

Temperature-SD-1000-1600 Capital

(4) (5) (6) The dependent variable is Culture-2008 0.191 (0.080)** 1.946 (0.940)*

- 0.140 (0.065)**

Culture-F-1000-1600 Coast

(3)

Catholicism Fixed country effects YES Estimation 2 R 0.17 Number of observations 87 Notes: 1. Standard errors clustered 2. The specifications always

- 0.134 (0.069)* 0.518 (0.249)* 1.290 (1.386)

- 0.143 (0.069)*

- 0.412 (0.337) YES OLS. 0.18 0.17 0.18 0.19 87 87 87 89 at the country level in parentheses. *** denotes significant include also Area, Latitude, and Longitude. YES

YES

YES

15

(7)

- 0.132 (0.067)*

0.090 (0.070) - 0.142 (0.069)*

- 0.462 (0.369) YES

- 0.421 (0.333) YES

(8)

- 0.139 (0.071)* 0.470 (0.308)

- 0.422 (0.340) YES

0.21 0.19 0.20 89 89 89 at the 1% confidence level; **, 5%; *, 10%.