key: cord-0877801-xrrly6t0
authors: Micheli, Martin
title: Aggregate stability under a budget rule and labor mobility()
date: 2020-09-01
journal: Econ Model
DOI: 10.1016/j.econmod.2020.08.010
sha: 1605bc38f1953f5b1473536b75c620a5989e10db
doc_id: 877801
cord_uid: xrrly6t0

Whether a balanced budget rule stabilizes or destabilizes an economy depends on various factors such as the production function or the instrument used to balance the budget. This paper argues that migration, which has widely been neglected in the literature, also affects equilibrium properties. We study the effect of pro-cyclical labor mobility in a neoclassical growth model with public debt and a balanced budget requirement. Labor mobility can destabilize the economy due to external effects. After a negative shock hits the economy, living abroad becomes relatively more attractive, resulting in out-migration. This increases per capita public debt as migrants leave behind their implicit liabilities. The government increases tax rates to satisfy the balanced budget requirement, which further depresses the economy and increases out-migration. The destabilizing effect of public debt kicks in at only slightly higher debt levels than the ones observed in the Euro area after the financial crisis.

Over the past 15 years, public debt levels in Europe have increased considerably. Prior to the financial crisis, average public debt in individual Euro area member states roughly equaled 70 percent of GDP. Due to measures to prevent a meltdown of the financial system during the financial crisis, budget deficits 5 to support the economy during the Great Recession, and the slow economic recovery especially in the periphery of the Euro area public debt increased to over 90 percent of GDP in 2014.

In 2020, public indebtedness relative to economic activity is expected to increase dramatically due to the COVID-19 pandemic. On the one hand, gov-10 ernments' revenues have been eroding as containment measures and behavioral changes of individuals pushed the economy into the most severe recession since World War II. On the other hand, governments' expenditures have been increasing due to expansionary fiscal policies to prevent the recession from being even more severe. The resulting increase in public indebtedness is especially trou-15 bling given the recent experience of the European Debt Crisis. Several member states had to pay severe premiums when borrowing or even lost access to financial markets due to doubts about their solvency. Via feedback loops between sovereign and banking sector insolvency, unfavorable financing conditions for governments transmitted to the financial sector of the respective countries [1] . 20 This paper argues that high levels of public debt might be even more worrisome than recognized in the literature. This holds especially true in Europe, where labor mobility has increased substantially over the past decades [2] . On the one hand, labor mobility has been identified as an important mechanism to cushion adverse demand shocks. Increasing labor mobility has therefore been an 25 explicit goal of European policymakers [3] . On the other hand, labor mobility in the presence of public indebtedness has external effects as emigrants leave behind implicit liabilities: per capita public debt.

The literature already discusses potential negative effects of high public indebtedness. The vigorous debate about the effect of public debt on economic 30 growth is probably the most prominent example. [4] argue that economic economic growth sharply declines if public debt reaches a threshold value of about 90 percent of GDP. [5] challenge this view. There seems to be a negative correlation between public debt and economic growth. The threshold for the sharp decline of economic growth, however, seems to be sensitive to the weighting 35 scheme and the sample period.

High public debt levels might also affect a government's ability to respond to economic shocks with counter-cyclical fiscal policy. Banking crisis typically last longer for highly indebted countries than for countries with ample fiscal space [6] . [7] show that countries with low public debt typically conduct counter-40 cyclical fiscal policy to dampen the business cycle. If, however, the stock of outstanding debt exceeds 86 percent of GDP, fiscal policy becomes pro-cyclical as debt sustainability becomes an issue. Using a general equilibrium model, [8] show that observed default premiums during the European Debt Crisis can be explained by the emergence of sovereign risk as governments reached their fiscal 45 limit.

To prevent unsustainable public debt levels, the European Union introduced limits to government debt and deficits for their member states in the Maastrich Treaty. As a response to the Great Recession and the European Debt Crisis, the Stability and Growth Pact has been reformed in the Sixpack with the aim of 50 ensuring sufficient fiscal space in future crises. The specific design of a budget rule has substantial implications for the welfare of such a rule [9] .

The theoretical literature typically discusses the effects of a balanced budget requirement on the stability of the equilibrium allocation. In a seminal contribution, [10] introduce a balanced-budget requirement into a neoclassical growth 55 model with distortionary taxation of labor income. Such a rule can result in aggregate instability for empirically plausible parameter constellations. Many subsequent studies show that aggregate stability in models with balanced-budget rules depends on whether the government adjusts the budget via the income or via expenditures [11] , the supply of public goods [12], the tax type [13] , the 60 utility function [14] , and the production function [15, 16, 17] . In a small open 2 J o u r n a l P r e -p r o o f economy, indeterminacy is less of an issue due to integration in international capital and goods markets [18, 19] . But even there, indeterminacy can occur depending on the level of production externalities and and the tax scheme [20] .

We contribute to the literature on sustainability of public debt by introduc-65 ing a widely neglected aspect: labor mobility. 1 The evolution of the population size since the Great Recession is another, admittedly purely descriptive, indicator that individuals take advantage of free movement within the Schengen area. We show that the interactions of labor mobility, public debt and the presence of a balanced budget requirement have implications for the stability of 80 the economy. Under a standard parameterization and public debt levels only slightly higher than the ones observed in the Euro area after the financial crisis, the model ceases to be saddle-path stable. The mechanism is the following:

A negative shock results in utility losses for domestic individuals. This makes living abroad relatively more attractive, resulting in out-migration. As public 85 debt per capita increases, the government has to adjust the tax rate, putting further drag on the economy, resulting in further out-migration. Our results are robust with regard to the assumption about capital mobility and different assumptions about fiscal policy.

We augment a real business cycle model with labor mobility. The economy is small relative to the rest of the world such that migration flows do not affect global variables.

The economy is populated by N infinitely-lived identical individuals. The 95 index i indicates that the respective variable is in per capita terms. The representative individual i supplies one unit of labor inelastically and maximizes expected lifetime utility

where E is the conditional expectations operator and β is the discount factor with 0 < β < 1. The representative individual derives utility from consumption 100 c it and faces the budget constraint 

Firms produce output according to a Cobb-Douglas production function.

We aggregate individual firms and describe the maximization problem of the production sector. All variables are therefore in aggregate terms. Aggregate production y t requires the two input factors labor h t , which each individual supplies inelastically, and working capital k t−1 that is decided on in period 115 t − 1. a t represents total factor productivity.

Firms maximize the sum of discounted expected profits given the discount

Accumulation of working capital takes place at the firm level. It is subject to the depreciation rate δ ∈ (0, 1). Besides working capital, firms buy government 120 bonds b t , which pay the risk free interest rate r b .

This setup results in the first order conditions

The government faces an initial stock of public debt b t−1 , which is held by domestic firms. Each period, the government has expenditures due to repaying 125 maturing public debt and due to wasteful government consumption g. Government consumption is proportional to the population size N t−1 . 2 Per capita government consumption g i depends on economic activity. The government raises income by taxing households' consumption expenditures and has income due to issuing debt that has to be repaid in the following period. The budget 130 constraint reads

c it represents aggregate consumption expenditures of households. The government stabilizes per capita debt such that

The literature typically abstracts from labor mobility and assumes that the population is constant. In such a setup, stabilizing debt and stabilizing per capita 135 debt are equivalent. This is not the case if the workforce is mobile. We assess stabilizing capita debt to be more sensible. The Maastricht treaty for example sets limits for the debt to gdp ratio. Gdp, however, is affected by the population size for obvious reasons.

The determinants of population size and the persistence of shocks are still a controversial topic in economic geography. There are mainly three competing theories [30]: Local fundamentals, random growth, and increasing returns to scale. Local fundamentals theory states that population size is determined by local fundamental factors such as the geography. Population size therefore follows a deterministic trend, temporary shocks only have temporary effects on population size. In the random growth theory, population growth follows a stochastic trend, temporary shocks therefore have lasting effects on population size. Increasing returns theory also postulates permanent effects of temporary shocks. As multiple equilibriums are inherent in this theory, temporary shocks 150 can shift the economy to a new equilibrium.

The empirical evidence on the drivers of population size is mixed.

[30] investigate the distribution of regional economic activity, which should correlate with population size, in Japan. The authors analyze its development over a long time period and after a temporary shock, the bombings of Hiroshima and 155 Nagasaki, and find evidence for a combination of the above theories. The spatial pattern seems to be determined by fundamentals. Quantitative differences in this pattern, however, are attributed to increasing returns. For Germany, city growth seems to depend on the political system [31] . The authors show that strategic bombings during World War II had permanent effects in East 160 Germany, while the effects in West Germany were only temporary. [32] uses variation in the absorption of expellees over the different occupational zones.

In this analysis, population patterns were highly persistent, fundamentals did not determine this pattern. The author argues that these results are consistent with empirical evidence on path dependence on regional economic development 165 [33, 34, 35] . This paper follows [32] by assuming that temporary shocks have permanent effects on population size. The process for population size exhibits a unit root.

Each individual decides in period t about the country of residence in period t + 1. This decision is based on a comparison of expected flow utility in the 170 home economy to an alternative abroad. Similar to [21] , population size follows a unit root process. 3

c * i represents consumption in case of living abroad. The home economy does not affect the world economy as it is assumed to be relatively small. Consumption abroad is therefore stable over time. 

Similar to [21] , our formulation of the migration function results in population size N to exhibit a unit root. Therefore, other variables such as consumption c and output y also are non-stationary. We can transform this nonstationary system to a stationary one by expressing the model in per capita 185 terms (relative to the current period's population size N t−1 ) and using population growth Π N t = N t /N t−1 instead of its size. Assuming that individuals and firms discount future utility and profits similarly, the system of equations 3 [21] assumes that migration flows depend on differences in lifetime labor income relative to an outside option. He employs a growth model without micro-foundation, is therefore not able to compare individual utilities. Other popular approaches to model the migration decision are to compare wages [27] or in search and matching models to compare expected returns from searching for a job [29] in the two countries. However, all modeling approaches represent shortcuts as the migration decision depends on a variety of factors [36] that are typically not accounted for in general equilibrium models.

4 Individuals decide in period t in which economy they want to live in period t + 1. This decision is based on current period's information, only. Therefore, Nt represents the domestic population in period t + 1, decided on in period t.

J o u r n a l P r e -p r o o f is given by

as well as a process for total factor productivity log (a t ) = log (a) (1 − ρ) + 190 ρ log (a t−1 ) + t , t represents a shock to total factor productivity with t ∼ NIID (0, σ 2 ), and a process for per capita government expenditures g it .

The model allows us to mute labor mobility by setting µ = 0. Note that muting labor mobility, which is equivalent to replacing (14) with the Π N t = 1, the model collapses to a rather standard rbc model. 5 Note that the production function (4) determines GDP per capita as y i = ak i α .

J o u r n a l P r e -p r o o f

The model is calibrated for an annual frequency. We follow the calibration used in the small open economy literature [37] . We set the capital share α = 0.32 and the discount factor β = 0.96, resulting in a steady state real interest rate of roughly 4 percent. The annual depreciation rate of capital is 10 percent 210 (δ = 0.1). We also use the authors' value for ρ = 0.42, the coefficient of auto-correlation of total factor productivity, and for the standard deviation of innovations to technology σ = 0.0129.

To show the dynamics of the model, we set the steady state government debt to GDP ratio θ b to 0. We fix primary per capita government expenditures 215 to GDP (θ g ) at 0.45 and assume that expenditures do not react to changes in output per capita. This allows us to show the effect of a shock to total factor productivity on equilibrium dynamics. As the main aim of this paper is to analyze the effect of government debt on aggregate stability, we will introduce government debt later on. A summary of the calibration is shown in Table 1 . 

In this section, we show equilibrium dynamics in the model with labor mobility. We present dynamics for five different values of µ, the percentage increase in the domestic population if individual consumption in the domestic economy is 1 percent higher than consumption abroad. The five different values are re-230 ported in Table 2 . As described in Section 2.7, we mute the effect public debt by assuming θ b = 0. 

Let us now investigate the effect of government debt, primary expenditures 250 and labor mobility on aggregate stability (Figure 3 ). Shaded areas indicate 6 We use the software package Dynare to simulate the model [38] .

J o u r n a l P r e -p r o o f To increase transparency of what is happening in the economy, we rewrite a deterministic system of the equations (10) to (14) in matrix notation as:

Per capita consumptionĉ i and the per capita capital stockk i are in percentage deviations from steady state. We abstract from exogenous shocks to total factor 270 productivity and define Φ 1 and Φ 2 as

to simplify the notation. This system has one forward looking variable, per capita consumptionĉ i , and one backward looking variable, the capital stock k i . For low values of µ, one eigenvector's absolute value is larger and one is 275 smaller than. The system is saddle-path stable. If, however, labor mobility increases beyond the thresholds presented e.g. in Figure 3 , the eigenvalue with the absolute value smaller than becomes larger than one. The transversality conditions cannot be satisfied and the system ceases to be saddle-path stable.

Sovereign default becomes inevitable. 280 

In this section, we use alternative parameter constellations to get an idea about the robustness of our results. In the following analysis, we set the elasticity of population size with respect to consumption µ to 0.2. 285 We first allow for government expenditures to depend on economic activity as suggested by [14] . More specifically, we assume g i (y it ) = g i Note: Shaded areas indicate parameter constellations without saddle-path stability. θ g and θ b represent the government share and the debt to gdp ratio.

Until now, we assumed that labor is mobile and capital is immobile. However, in the open economy literature, capital is mobile and capital flows ensure the equalization of interest rates. In this section, we therefore allow for capital to move cross country borders.

Firms accumulate foreign debt s. The production sector's budget constraint now is

It is well known that assuming an exogenous interest rate, which is given by the world interest rate r b t = r rf , results in instationarity. We follow [37] and induce stationary by assuming that the interest rate depends on per capita foreign debt.

This results in the additional equation

with r rf being the exogenous world interest rate. s i is steady state per capita foreign debt. We set ψ to 0.0742.

Equilibrium dynamics in a model with capital mobility are very similar to the ones in the model without capital mobility ( Figure A2 in the appendix).

Again, the stronger migration reacts to differences in consumption, the less pronounced are the effects of shocks. 

We now analyze the effect of foreign debt on aggregate stability. For positive values of the debt to GDP ratio θ s , the economy is a net debtor. For negative values, the economy is a net lender to the world economy. Figure 6 shows that saddle-path stability increases with foreign assets. The intuition is Note: Shaded areas indicate parameter combinations without a stable steady state. θ g and θ b represent the government share and the debt to gdp ratio.

straightforward. As foreign assets and debt are accumulated by the production sector, migration flows not only affect per capita government debt but also the per capita foreign investment position. Out-migration therefore increases per capita debt. In case of foreign assets, out-migration increases the per capita international investments, which counteracts the effect of government debt. 325 

The strong dependence of the results on µ, the effect of differences in consumption opportunities on migration, raises the question of the relevance of the proposed channel of government debt on aggregate stability. To get an idea about the value of µ, we estimate the equation

Using π N = log Π N , (18) represents the logarithmic approximation of (9)

augmented by a constant and an error term. We estimate this equation for the This leaves us with a decision on the outside option. We assess average per capita consumption in the Euro Area and in the European Union so be equally plausible. We therefore report the results for both specifications. The evolution 340 of all the variables is shown in Figure A1 in the Appendix. All information is available at Eurostat. Note: * p < 0.10, * * p < 0.05, * * * p < 0.01.

The correlation seems to be positive and statistically significant (Table 3) .

However, we want to emphasize that the purpose of this paper is to point out the potential channel, not to estimate a causal effect of differences in real con-345 sumption on population growth. The estimation is meant to give an idea of whether this channel could have any practical relevance.

Given these estimates, our simple model would suggest that migration flows

are not yet at levels that result in an economy characterized by the absence of saddle-path stability. This can easily seen by comparing estimated coefficients 350 with the parameter space with saddle-path stability of e.g. Figure 3 . Further increases in public debt that are almost certain given the severity of the current recession, however, might represent an actual danger to saddle-path stability.

There are several shortcomings to our calculations. Our estimates are subject to endogeneity, as changes in per capita private consumption expenditures also 355 affects per capita consumption in the country aggregate. We asses this effect to be negligible, especially for Cyprus, Greece, Ireland, and Portugal due to their size. One could also argue that the EMU and even the EU does not represent the universe of countries of origin for immigrants and of destination countries for emigrants. We introduce labor mobility into a simple real business cycle model with public indebtedness and a balanced budget requirement for the government. In such a setup, government debt decreases the parameter space of saddle-path 370 stability as migration has external effects when the government is indebted.

Individuals leaving the economy shift the burden of debt service to the remaining population. Per capita public debt service is counter-cyclical. This way, labor mobility can push otherwise saddle-path stable economies to unstable ones.

Non-existence of a saddle-path stable equilibrium can occur for government 375 debt levels only slightly higher than the ones observed in the Euro area after the financial crisis.

We therefore argue that reducing public indebtedness of individual member states when the economy has recovered from the COVID-19 pandemic is vital for stability in the European Union. This is especially true as increasing 380 labor mobility is a political goal in the EU. We further argue that migration flows should be endogenized in medium term projections of the European Commission. The current commonly agreed method for calculation potential output treats the population size as exogenous. Pro-cyclicality of migration flows might therefore results in an underestimation of potential risks of high levels of public 385 debt.

Deadly Embrace: Sovereign and Financial Balance Sheets Doom Loops

Labour market adjustments and migration in Europe and the United States: How different?, Economic Policy

Labour mobility in the EU, speach at the University of Ghent

Public debt overhangs: Advanced-economy episodes since 1800

Does high public debt consistently stifle economic growth? A critique of Reinhart and Rogoff

How effective is fiscal policy response in financial crises

Is fiscal policy always counter-(pro-) 405 cyclical? The role of public debt and fiscal rules

Default risk premia on government bonds in a quantitative macroeconomic model

Does the design of a fiscal rule matter for welfare?

Balanced-budget rules, distortionary taxes, 415 and aggregate instability

Balanced-budget rules and macroeconomic (in)stability

Transitional dynamics and indeterminacy of equilibria in an endogenous growth model with a public input

Balanced budget rules and aggregate instability: The role 425 of consumption taxes

Aggregate instability under balanced-budget consumption taxes: A re-examination

Destabilizing balanced-budget consumption taxes in multi-sector economies

Aggregate stability and balanced-budget rules

Public debt and aggregate stability with endogenous growth and a state-dependent consumption tax

Balanced-budget consumption taxes and aggregate sta-440 bility in a small open economy

Balanced-budget income taxes and aggregate stability in a small open economy

Stabilization policy and indeterminacy in a small open economy

How does labor mobility affect income convergence?

Migration dynamics, growth and convergence

The US westward expansion

Migration selectivity and the evolution of spatial inequality

Productivity differences and the dynamic effects of labor movements

Immigration, Search and Redistribution: A Quantitative Assessment of Native Welfare

Immigration, remittances and business cycles

A search and matching approach to business-cycle migration in the euro area, no. B06-V3 in Beiträge zur Jahrestagung des Vereins für Socialpolitik 2019: 30 Jahre Mauerfall -Demokratie und Mark-470 twirtschaft -Session: Macroeconomics -Labor Markets

Economic migration and business cycles in a small open economy with matching frictions

Bones, bombs, and break points: The geography of economic activity

The strategic bombing of German cities during World War II and its impact on city growth

Persistence of population shocks: Evidence from the occupation of West Germany after World War II

History and industry location: Ev-485 idence from German airports

Portage and path dependence

Local economic development, agglomeration 490 economies, and the big push: 100 years of evidence from the Tennessee valley authority

In-495 ternational Migration Drivers, Publications Office of the European Union

Closing small open economy models

Measuring cyclically-adjusted budget balances for OECD countries

Adjusting the budget balance for the business cycle: The EU methodology, Tech. rep., Directorate General Economic and Financial Affairs (DG ECFIN)

 Figure A1 : Population growth and relative consumption opportunities