Does Competition for Capital Discipline Governments? Decentralization, Globalization, and Public Policy

Theoretical Foundations of Competition for Capital

• The Question of Discipline: The central inquiry asks whether the competition to attract mobile capital acts as a disciplining force on governments. Two distinct bodies of literature address this, both agreeing that the discipline exists but disagreeing on its normative outcomes.

• The "Harmful Discipline" School:     * Scholars in this camp argue that the fear of capital outflows constrains governments from providing essential public goods such as welfare services, environmental regulations, and "nonproductive" public goods valued by citizens.     * Wallace Oates (1972): Stated that local officials may keep taxes low to attract business, holding spending below levels where marginal benefits equal marginal costs, especially for programs without direct business benefits.     * George R. Zodrow and Peter Mieszkowski (1986): Modeled the mechanism of this under-provision.     * Michael Keen and Maurice Marchand (1996): Argued capital competition distorts spending toward infrastructure (business centers/airports) and away from other goods (parks/libraries).     * John H. Cumberland (1981): Contended that inter-jurisdictional competition weakens environmental standards.     * Mark C. Rom et al. (1998): Discussed a "race to the bottom" in U.S. welfare and social services.     * Dani Rodrik (1997): Argued capital mobility makes it harder to provide social insurance.     * Schulze and Ursprung (1999): Suggested states restructure expenditure toward productive public inputs at the expense of transfers.

• The "Salutary Discipline" School:     * This group views discipline as beneficial, motivating governments to reduce corruption, waste, and inefficiency.     * Yingyi Qian and Grard Roland (1998): Argued interregional competition punishes wasteful governments with capital flight.     * Gabriella Montinola et al. (1995): Suggested competition induces governments to provide hospitable environments, secure property rights, and infrastructure. They cite China as an example where competition led to pro-business laws.     * Maurice Obstfeld (1998): Contended international capital markets discipline policymakers tempted to exploit captive domestic capital; unsound policies (e.g., inadequate bank regulation) spark outflows.     * Joseph E. Stiglitz (2000): Acknowledged that opening capital accounts "forces" countries to have good economic policies.     * The Economist (2001): Famously stated that integration makes it "harder to be a tyrant" because people can leave with their savings.

The Authors' Core Thesis: Heterogeneity and Polarization

• The Flaw in Previous Models: Cai and Treisman argue that the discipline effect is not as general as previously assumed because standard models rely on the assumption that units (regions or countries) are identical (symmetric equilibria).

• The Endowment Hypothesis: If some units start with better "endowments" (natural resources, geographical advantages, or inherited human capital), symmetric equilibria do not exist.

• The Polarization Effect: * If endowment differences are sufficiently large, poorly endowed units will have less business-friendly policy under capital mobility than under immobility. * Better-endowed units (e.g., New York, Moscow, East Asian industrial zones) will invest heavily in business services and pull capital away from poorly endowed counterparts. * Poorly endowed units (e.g., Chad, the Russian republic of Buryatia) realize they cannot compete even with infrastructure investment. Consequently, they give up on pro-business policies and focus on predation or satisfying local demands, facing less effective discipline.

Mathematical Model: Capital Immobility vs. Mobility

• Production Technology: The model assumes a Cobb-Douglas production function for unit $i$:     $F_i = A_i k_i^\alpha I_i^\beta$     * Variables: $A_i$ (exogenous endowment effect on output), $k_i$ (capital), $I_i$ (infrastructure investment).     * Constraints: \alpha > 0, \beta > 0, \alpha + \beta < 1 (accounting for fixed factors like land/labor).     * Asymmetry: $A_n$ (well-endowed units) vs. $A_m$ (poorly endowed units), where A_n > A_m.

• Government Payoff Function:     $U_i = (1 - t_i)F_i + \lambda c_i$     * Variables: $t_i$ (tax rate), $c_i$ (government spending/consumption), $\lambda \geq 0$ (preference for public spending relative to private consumption).     * Budget Constraint: $I_i + c_i = S + t F_i$.

• Case A: Capital Immobility:     * Capital levels ($k_n, k_m$) are trapped. The government maximizes $U_i$ subject to its budget.     * The first-order condition (FOC):     $\frac{\partial F_i}{\partial I_i} = \gamma$     * Where $\gamma = \frac{\lambda}{1 + (\lambda - 1)t}$. This represents the opportunity cost of infrastructure investment.     * Result: $I_i(k_i, A_i) = [(\frac{1}{\gamma}) \beta A_i k_i^\alpha]^{\frac{1}{1 - \beta}}$. Investment increases with capital and endowment levels.

• Case B: Capital Mobility:     * Capital flows until net returns are equalized: $(1 - t) \frac{\partial F_i}{\partial k_i} = r$.     * Capital allocation rule: $k_i(I_i, r, A_i) = [\frac{(1 - t) \alpha A_i I_i^\beta}{r}]^{\frac{1}{1 - \alpha}}$.     * The FOC for the government under mobility:     $\frac{\partial F_i}{\partial I_i} + (\frac{\partial F_i}{\partial k_i})(\frac{\partial k_i}{\partial I_i}) = \gamma$     * The second term represents the indirect effect of infrastructure attracting additional capital.

• Result of Competition vs. Polarization:     * Competition Effect: Units invest more to attract capital.     * Polarization Effect: A shift in capital allocation towards well-endowed units.     * As $\frac{A_n}{A_m}$ increases, $k_m$ and $I_m$ decrease. For large endowment gaps, poorly endowed units invest less than they would if capital were immobile.

Extensions and Robustness

• General Production Functions: The conclusions hold for any increasing and concave output function $f(k_i, I_i, A_i)$. High asymmetry can lead to a "total polarization equilibrium" where infrastructure and capital in poorly endowed units drop to zero.

• Quasilinear Payoff Functions: Replacing the linear utility with $U_i = (1 - t)F_i + \lambda v(c_i)$ (where v' > 0, v'' < 0) maintains the qualitative results.

• Strategic Interaction: In cases with a small number of units, governments anticipate that their investments affect the economy-wide rate of return $r(I_i, I_{-i})$. This typically strengthens the polarization effect because infrastructure increases by rivals raise the general return, making local investment less effective.

• Endogenous Tax Competition: When governments set both tax rates and infrastructure, the disciplining effect is even weaker. Higher taxes needed for infrastructure discourage investors:     $\frac{\partial F_i}{\partial I_i} + (\frac{\partial F_i}{\partial k_i})(\frac{\partial k_i}{\partial I_i}) = \gamma + \lambda v'(c_i) \times [- \frac{(1 - t_i)(\frac{\partial F_i}{\partial k_i})(\frac{\partial k_i}{\partial t_i})}{F_i}]$

Empirical Illustrations: Russia and the Developing World

• Post-Communist Russia:     * Following 1991, capital moved freely. Rather than a general improvement, a split occurred between well-endowed regions (Moscow, St. Petersburg, Samara) and poorly endowed ones (Altai, Tyva, Kalmykia).     * Initial endowment index (natural resources, geography, human capital, infrastructure) correlated positively with:         * The share of regional budgets spent on transport, roads, and market development (r = 0.48, p < 0.01).         * Ekspert magazine ratings for market economy institutions as of 2001 (r = 0.41, p < 0.01).     * Investment patterns: Well-endowed regions with business-friendly policies saw significant net capital inflows. Poorly endowed regions let physical infrastructure decay and suffered net outflows of local savings.

• Capital Account Liberalization in the Developing World:     * Developing world share in global private capital flows fell from $11.8\%$ in 1991 to $7.6\%$ in 2000, despite output growth.     * Inflows were concentrated in a few countries (China, Mexico, Brazil).     * Sub-Saharan Africa: Private capital inflows fell from $3.9\%$ of GNP (1975–1982) to $1.8\%$ in 1990–1998.     * Capital flight: In countries like Egypt, Mauritius, and Uganda, outflows by residents rose after liberalization. By the late 1990s, Africans held more wealth overseas than residents of any other continent.     * Infrastructure Lag: African rural highways (55 km per 1,000 sq km) remain far behind India (800 km). Paved roads in Africa actually decreased during the 1990s.

Conclusions and Policy Implications

• Summary of Findings: Capital flow is not a universal disciplining force. When countries differ markedly, better-endowed units drain capital, and poorly-endowed units "give up."

• Policy Suggestions for Exploration:     1. External Aid: Directing aid to market infrastructure or human capital in poor regions may reduce the gap to the point where competition discipline finally activates.     2. Regional Clubs: Liberalizing capital within blocks of similar endowments (e.g., within the EU or specifically within African states) may prevent immediate capital drain to the global market.     3. Decentralization Caveat: Decentralization may discipline governments in homogeneous countries but may exacerbate bad governance in geographically/economically diverse ones.