Property Rights, Land Liquidity and Internal Migration

Property Rights, Land Liquidity and Internal Migration By Eugenia Chernina, Paul Castañeda Dower and Andrei Markevich November 15, 2011 LLMS&CLMS seminar, Higher School of Economics

Hypothesis The 1906 Stolypin titling reform had a positive impact on internal migration from European to Asian part of Russian Empire through the following mechanisms: improved land liquidity: possibility of land sale, mortgage or rent eased financial constraint of potential migrants. improved tenure security: reduced opportunity costs of migration

Presentation plan Relation to literature Historical background The simple model of migration Empirical strategy Empirical results Conclusions

Relation to Literature Imperfect credit or insurance markets might initiate migration: Stark 1991 Wage differentials and migration: gravity models Evidence of financial constraints for migration: Hatton and Williamson 1994, Andrienko and Guriev 2004, Halliday 2006 Impact of property rights improvements onto migration through tenure security: Field 2007, Do and Iyer 2008, De la Rupelle et. al. 2009 Discussion of Stolypin reforms in Russian historical literature: Dubrovski 1963, Zyrianov 1992

The 1906 Reform as treatment Before the reform After November, 9 1906 decree Repartion commune Communal rights. Can not sell their plots. Right to exit commune, convert communal title to hereditary tenure even without consent of the commune and with legal right to sell, lease or mortgage the plot Heridatory commune Personal rights, quasi selling was possible Almost not affected

Dynamics of title conversions 700000 600000 500000 400000 300000 200000 100000 0 1907 1908 1909 1910 1911 1912 1913 1914 1915

What else happened in 1906? 10 th of March, 1906 decree introduced new migration policy: enlarged existing subsidies for migrants interest-free loans for socially valued projects special governmental programs on land research, demarcation and improvement development of public health in the regions of destinations a wide information campaign on migration to Siberia simplified procedure to get official permits to migrate

Migration dynamics 1901-1914

A Simple Model of Migration xip ~F(x; µ, σ) - wealth of household i in province p λp - province (il)liquidity parameter Cp - the costs of migration Sip - subsidies ξip = (Cp - Sip)/(1-λp) budget constraint νmig - expected benefit of migrating γip - tenure security parameter v( x ip, γ ip v ) = Mig [1 ( δ δ 1 δ R R )] ( γ ipxip + ϕ) δ + 1 δ Is increased by reform in repartition communes Are increased simultaneously for all provinces Is increased by reform in repartition communes Pm = F(ν > x ξ; µ, σ) - share of migrating population

Hypotheses Land Liquidity Prediction: The proportion of migrants in a province is negatively related to the illiquidity of land. In particular, F(ν > x ξ; µ, σ)/ λ< 0. Tenure Security Prediction: The proportion of migrants in a province is positively related to the strength of tenure security. In particular, F(ν > x ξ; µ, σ)/ γ> 0. The higher is this effect - the higher is temporary migration.

Models with Reform Dummy Difference-in-Difference Model: Migrationit = α + β*repartitioni + d* RPeriodt + (1) γ*repartition*rperiodit+φxit + εit Fixed effects model: Migrationit = α + γ*repartition*rperiodit + φxit + ui + τt + εit (2) Migrationit - number of migrants per a thousand of citizens. Repartitioni - dummy indicating provinces affected by the reform RPeriodt dummy for the date of implementation of the reform Xit is a set of control variables: rural density, average titled plot size, unskilled wage rate in a province, rural wages, livestock, seed yield per square kilometer, railway tariffs ui and τt - provincial and time fixed effects εit - random disturbance

Models with Exits from Commune Migrationit = α + β*number of exitsit + φxit +ui + τt + εit (3) Number of exits number of households who performed an exit from the commune accompanied by title conversion Inverse causality is a concern: upward bias is possible because of migrants who would anyway migrate Proposed instrument: number of exits could be instrumented by the proportion of applications over title conversion that were quickly confirmed by the local authorities (confirmation_rate). The variable is interpreted as a measure of bureaucratic quality. Thus, first stage estimator is: Number of exitsit=α + β*[confirmation_rate]it + φxit + ui + τt+ ηit (4)

Data Sources Data has period/province structure Observation periods are: 1901-1906, 1907, 1908-1909, 1910-1911, 1912, 1913 and 1914 Migration statistics for 1901-1914 by Resettlement Administration: subsidized and unsubsidized migrants registered when passing key railway stations (Syzran and Chelyabinsk) Peasants exit applications statistics for 1907-1915 by the Minestry of Internal Affairs published in Izvestiya zemskogo otdela. Peasants sales of privatized plots in repartition communes from annual statistical volumes of the Ministry of Justice Population, rural density, urbanization, yield and livestock statistics from official statistical volumes annually published by the Central Statistical Committee of the imperial Ministry of Internal Affairs since 1904 Rural wages from agricultural statistical volumes published annually by the Ministry of Agriculture since 1906 Urban wages from annual labor inspection reports published by the Ministry of Finance Railway tariffs from a 1911 information booklet published by the Resettlement Administration for migrants.

Results for Models with Reform Dummy Dependent Variable: Migration per capita Diff-n-Diff Pooled OLS Fixed Effects 1 2 3 4 repartition_rperiod 0.20*** 0.18*** 0.15*** 0.19*** [0.050] [0.053] [0.050] [0.062] rperiod 0.11*** 0.14** 0.05 [0.032] [0.067] [0.069] repartition_province 0.04* 0.11 0.22** [0.023] [0.082] [0.100] Controls No Yes Yes Yes Year Effects No No No Yes Regional Effects No No Yes No Provincial Effects No No No Yes Constant 0.06*** 0.02 0.94*** 1.57*** [0.016] [0.264] [0.274] [0.558] Observations 349 340 340 340 R-squared 0.082 0.235 0.396 0.378 Number of id 50

Results for Models with Exit Number Dependent variable Migrants per capita Exits Migrants per capita Pooled OLS Fixed Effects IV Fixed Effects Fixed IV Fixed First Stage Effects Effects VARIABLES 1 2 3 4 5 6 7 Exits 0.04*** 0.04*** 0.03*** 0.03** 0.03*** 0.03*** Repartition_province 0.16** 0.19** [0.007] [0.008] [0.008] [0.013] [0.008] [0.011] [0.071] [0.087] Land Sales 0.05* 0.05* Repartition_rperiod 0.07 0.05 Rperiod 0.14** [0.054] [0.058] [0.066] Confrate 17.74*** [0.025] [0.024] [2.844] Controls Yes Yes Yes Yes Yes Yes Yes Time Effects No Yes Yes Yes Yes Yes Yes Regional Effects No Yes No No No No No Provincial Effects No No Yes Yes Yes Yes Yes Constant -0.40 0.52** 2.28*** 0.45 2.27*** [0.277] [0.205] [0.516] [4.357] [0.520] Observations 319 319 319 319 318 318 317 R-squared 0.374 0.586 0.474 0.707 0.466 0.482 0.474 Numberofid 47 47 47 47 47

Testing for mechanism Long-term Migrants Temporary Migrants Short-term Migrants All Migrating Hhs All Migrating Hhs Subsidized Migrating HHs All Migrating Hhs All Migrating Hhs (1) (2) (3) (4) (5) exits 0.03*** 0.00*** 0.00 [0.007] [0.000] [0.000] Lag1exits 0.00** 0.00** [0.001] [0.001] Lag2exits 0.00 [0.001] Controls Yes Yes Yes Yes Yield, Rwage, Urbanwage Constant 2.04*** 0.10*** 0.07*** -0.01 0.14 [0.456] [0.027] [0.016] [0.070] [0.120] Time Effects Yes Yes Yes Yes Yes ProvinceEffects Yes Yes Yes Yes Yes N of obs 310 311 311 268 226 R-squared 0.446 0.391 0.174 0.369 0.334 N of id 47 47 47 47 47

Dependent Variabl e Results for Subsidized and Unsubsidized Subsidized Migrants per capita Unsubsidized Migrants per capita Subsidized Migrants per capita Unsubsidized Migrants per capita Fixed IV Fixed First Stage, OLS OLS OLS OLS Fixed Effects IV Fixed Effects Effect s Effec ts Fixed Effects IV Fixed Effects 1 2 3 4 6 7 8 9 10 11 IV Fixed Effects(both IVs) Exits 0.02* 0.02 0.02** 0.02** 0.03*** 0.03*** [0.008] [0.010] [0.007] [0.006] [0.005] [0.005] repartition_r period 0.06 0.09* 0.09*** 0.10*** 0.02 0.01 3.61*** [0.037] [0.053] [0.019] [0.016] [0.045] [0.048] [0.511] rperiod 0.05 0.00 [0.045] [0.032] repartition_p rovince 0.21** 0.01 [0.080] [0.034] Controls Yes Yes Yes Yes Yes Yes Yes Yes Yes Yes Yes Time Effects No Yes No Yes Yes Yes Yes Yes Yes Yes Yes Regional Yes Effects No Yes No No No No No No No No Provincial No Effects Yes No Yes Yes Yes Yes Yes Yes Yes Yes Constant 0.47** 1.00*** 0.47*** 0.57* 1.45*** 0.83*** 4.50 [0.182] [0.333] [0.143] [0.299] [0.331] [0.286] [4.606] Hansen s J Statistic 2.88 (p-value) (.09) Observations 340 340 340 340 319 318 319 318 320 319 318 R-squared 0.425 0.311 0.264 0.366 0.360 0.345 0.470 0.470 0.524 0.451 0.445 Number of id 50 50 47 47 47 47 47 47 47

Robustness check: alternative treatment and control groups All Migrating Hhs Excluding Baltic Provinces Unsubsidized Migrating Hhs Expanded Control Group <30% repartition commune + Cossack provinces Unsubsidized All Migrating Hhs Migrating Hhs repartition_reform 0.11 0.09*** 0.11* 0.08*** [0.076] [0.025] [0.060] [0.024] rperiod 0.07-0.00 0.08 0.01 [0.098] [0.043] [0.072] [0.031] repartition_prov 0.25** 0.01 0.14 0.07 [0.114] [0.040] [0.095] [0.040] controls Yes Yes Yes Yes Regional Effects Yes Yes Yes Yes Constant 0.98*** 0.50*** 0.77** 0.25 [0.309] [0.154] [0.369] [0.155] Observations 307 307 340 340 R-squared 0.378 0.240 0.397 0.284

Robustness check: placebo treatment Subsidized Migrating Hhs All Migrating Hhs Subsidized Migrating Hhs All Migrants All Migrating HHs VARIABLES Pooled OLS Pooled OLS Pooled OLS Pooled OLS Pooled OLS 1 2 3 4 5 zemstvo_rperiod -0.02-0.09 [0.126] [0.109] SU_rperiod 0.24*** 0.16*** [0.081] [0.061] NTArperiod 0.04 [0.074] Rperiod 0.17 0.15 0.10 0.06 0.13 [0.127] [0.106] [0.068] [0.047] [0.093] Zemstvo -0.28** -0.23*** [0.113] [0.083] SU 0.02-0.04 [0.093] [0.058] NTransatlantic -0.05 [0.102] Controls Yes Yes Yes Yes Yes Time Effects No No No No No Region Effects Yes Yes Yes Yes Yes Constant 1.67*** 1.13*** 1.16*** 0.70*** 1.29*** [0.382] [0.286] [0.293] [0.205] [0.359] Observations 340 340 340 340 340 R-squared 0.399 0.455 0.399 0.417 0.378

Conclusions The initial hypothesis is approved: greater land liquidity due to the Stolypin reform promoted migration by easing financial constraints and decreasing opportunity costs. Empirical finding is that Stolypin Agrarian reform explains at least 18.1% of internal Europe-Asia migration in the late Russian empire. On average, for each 1000 title conversions from communal tenure, about 40 households subsequently migrate Contribution to the literature: Land liquidity turns out to be an important factor of migration decision; institutions improvement might promote rather than decrease migrants outflow Introduction of individual land titles improves the allocation of resources by influencing the decision to migrate Better understanding of the nature and design of the Stolypin agrarian policy by demonstrating that its two major components, namely the land titling reform and the encouragement of migration, were consistent with each other