Interest Premium, Sudden Stop, and Adjustment in a Small Open Economy

Abstract

This article studies the adjustment process of a small open economy to a sudden worsening of external conditions. The sudden stop is modeled by the use of a highly nonlinear specification that captures credit constraints in a convenient way. The advantage of this approach is that the effects of the shock become highly conditional on the external debt position of the economy. A two-sector model with money-in-the-utility is adopted, thereby making it possible to study sectoral asymmetries in the adjustment process, and also the role of currency mismatch. The model is calibrated to the behavior of the Hungarian economy in the 2000s, and its crisis experience in 2008–11 in particular. Four counterfactuals are calculated: two with different exchange rate policies (a more flexible float and a perfect peg), and both of these policy regimes with smaller initial indebtedness. Overall, the model is able to fit the movements of key aggregate and sectoral macroeconomic variables after the crisis by producing a large and protracted deleveraging process. It also offers a meaningful quantification of the policy tradeoff between facilitating the real adjustment by letting the currency depreciate and protecting consumption expenditures by limiting the adverse effect of exchange rate movements on household balance sheets.

Keywords:

• interest premium
• small open economy
• sudden stop

JEL Classification:

• E21
• E41
• E5
• F3

Notes

1. Slovenia was already on the euro in 2008. Bulgaria also had a large share of euro loans, but had a hard peg to the euro.

2. In order to match the initial developments in Hungary and the CEE region, a second, one-period shock is added that captures the large drop in foreign demand. While this may ultimately be caused by the same worldwide tightening of credit conditions, in the small open economy setting it is sufficient to implement it as a decline in foreign demand. Since this shock lasts only for one period, it plays essentially no role in the persistence of the effects of the crisis in the model economy, which is explained solely by the permanent shock to foreign borrowing conditions.

3. These properties make it very similar to the penalty function approach of Judd (1998), advocated recently by De Wind (2008) and Den Haan and Ocaktan (2009).

4. Note that the subscript j indexes investment targeted toward the accumulation of capital in sector j = X, NT, while the superscripts indicate the tradable (M) and nontradable components of these investments.

5. Any interest revenue would be rebated to households by the central bank.

6. Central banks could and do hold interest-bearing foreign assets. In the crisis period, however, interest earned on safe foreign assets, such as U.S. or German government securities, was essentially zero. Thus no distinction is made between foreign cash and other securities, but the analysis can easily be extended to take into account a more general foreign reserve composition.

7. In the interpretation and calibration the household sector also includes public debt, and government consumption and investment.

8. The local currency interest rate is defined as $R_t^d=R_t{S}_{t+1}/S_t$ using the uncovered interest parity (UIP) condition. Benczúr and Kónya (2013) show that under a flexible exchange rate and a constant money supply, nominal spending P_tC_t is constant. Combining this, the UIP equation and Equation (9) yield the desired result.

9. The main idea is to keep the precrisis ratio of reserves to M2 but allow for more exchange rate flexibility. A much higher ρ_s would produce excessive exchange rate depreciation, a huge initial drop in the NFA-to-GDP ratio, which in turn would prohibit Dynare from being able to solve the nonlinear model. A “true” flexible regime without less (or, in particular. zero) reserves would allow for a windfall NFA reduction through the sale of reserves, which is viewed as an irrelevant policy alternative.

10. Labor hoarding and government policies (like tax changes) could be partly behind the absence of a large employment drop. A decline in capacity utilization could also be an explanation. There is indeed an approximately 15% drop in the Eurostat series for Hungarian capacity utilization. This is large but still not sufficient to imply an unchanged employment level. From a growth accounting perspective, it leaves an unexplained drop in TFP

11. Detailed results for the two extensions described below are available from the authors upon request.

12. The shock itself is 485 basis points. After the endogenous response of all variables, the interest rate increases by 300 basis points. The (adjusted) steady state NFA per GDP position shifts by 1.01.

13. This exercise shows that there is a close mapping between the formulation of monetary policy (exchange rate smoothing) and a monetary reaction function responding to changes in the foreign currency premium.

Additional information

Notes on contributors

Peter Benczur

Peter Benczur is a Research Officer at the Joint Research Centre of the European Commission.

Istvan Konya

Istvan Konya is a Senior Research Fellow at the Centre for Economic and Regional Studies of the Hungarian Academy of Sciences, and an adjunct faculty at Central European University.