Italy: Selected Issues

International Monetary Fund

doi:10.5089/9781451819885.002.A004

Jump to Content Jump to Main Navigation

Italy: Selected Issues

Author: International Monetary Fund

Publication Date:: 16 Feb 2006

eISBN:: 9781451819885

Language:: English

Keywords:: ISCR; CR; export; private sector; factors accounting; infrastructure investment; income equation; tax multiplier; revenue shock; taxes rise; coefficient restriction; Disposable income; Real interest rates

Download PDF (1,007.1 KB)

This Selected Issues paper investigates the factors behind the deterioration in Italy’s international competitiveness. It concludes that the loss of competitiveness accumulated by Italian firms in recent years is mainly a consequence of weak long-term productivity performance. The paper explores the link between policies and growth. Specifically, it finds evidence that rigid product markets and a high tax burden on labor have been associated with slower growth in European regions. The paper also analyzes the role of fiscal policy and its implications for household consumption decisions.

Abstract

IV. How Expansionary Are Tax Cuts in Italy?¹

Core Questions, Issues, and Findings

What is the aim of the chapter? The chapter provides estimates of fiscal multipliers based on an intertemporal model for nondurable consumption where households are myopic, discounting the future at a rate higher than the prevailing real rate of interest. In such a theoretical framework, the impact of any shock to income/net taxes on consumption depends on three characteristics—the persistence of the shock, whether it is anticipated or not, and the discount wedge, i.e., the consumers’ excess of discount with respect to the market rate.
What are the main results of the analysis? The model entails a discount wedge of about 6 percent a year and fairly small fiscal multipliers of 0.05–0.2—depending on the permanence of the change in taxes/transfers. Historically, shocks to the net tax rate have been extremely short-lived, implying point estimates in the low end of the range. Accordingly, it seems improbable that changes in net tax rates would have significant effects on private consumption and, thereby, on growth.
What are the policy implications of this chapter’s findings? Thinking of fiscal policy in an intertemporal setting provides a range of insights. For example, to the extent that automatic stabilizers associated with the economic cycle are less persistent than other policy changes, they will be associated to smaller multipliers. Similarly, while it is unlikely that large fiscal contractions could be expansionary due purely to supply effects, it is possible that reductions in the real interest rate and changes in the assumed longevity of future tax cuts could also play a role. Finally, the implied low discount wedge suggests that households might cushion expected effects of pension reforms on life-cycle consumption by engaging in greater accumulation of assets.

A. Introduction

1. Fiscal policy remains an important lever for macroeconomic stabilization. The main issue associated with its effectiveness is the degree of Ricardian equivalence.² Full Ricardian equivalence implies that changes in taxes and transfers have no impact on rational consumers spanning an infinite lifetime. This is because optimizing agents discount the future using the interest rate on government paper, so the value of tax cuts and of subsequent tax increases exactly offset each other. Thus, rational consumers will fully offset a tax cut by increasing their saving.

2. There are two main ways of creating more realistic short-term tax multipliers. One is to assume that some individuals act as if they do not have access to financial markets, varying their consumption in line with their disposable incomes. The behavior of such credit constrained consumers is, however, highly mechanical, responding as much to a temporary tax cut as to a long-term one. The alternative is to assume that consumers have finite lives, adding a life-cycle dimension to consumption. This provides more realistic consumption dynamics, with spending responding less to a temporary tax cut than to a long-term one, as predicted by the permanent income hypothesis. In addition, the supply-side effects of fiscal policy can be incorporated by adding distorting taxes.³

3. The chapter provides estimates of fiscal multipliers for Italy based on an intertemporal model for nondurable consumption where households are myopic, discounting the future at a higher rate than the prevailing real rate of interest. In such a theoretical framework, the impact of any shock to income/net taxes on consumption depends on three characteristics—the persistence of the shock, whether it is anticipated or not, and the discount wedge, i.e., the consumers’ excess of discount with respect to the market interest rate.

4. The model entails a discount wedge of about 6 percent a year and fairly small fiscal multipliers of 0.05–0.2—depending on the permanence of the change in taxes/transfers. Historically, shocks to the net tax rate have been extremely short-lived, implying point estimates at the low end of the range. Accordingly, it seems improbable that changes in net tax rates would have significant effects on private consumption and, thereby, on growth. Results seem to be consistent with previous empirical evidence for Italy, indicating that net revenue shocks are very short-lived and tend to have negligible—both statistically and economically—effects on other macroeconomic variables.⁴

5. The chapter is organized as follows. Section B provides the theoretical framework for the analysis. The intertemporal model is estimated in section C. Section D explores these interactions in more detail. The low discount wedge implied by the estimates has, however, broader implications for policy analysis, which are briefly discussed in the concluding section.

B. Some Theory

6. In spite of practitioners’ renewed interest in the use of fiscal policy for macroeconomic objectives, research on fiscal multipliers remains limited compared to the empirical literature on monetary policy. To be sure, macroeconomic modeling groups continue to provide estimates of such multipliers, including in the context of a new breed of theoretically consistent “stochastic general equilibrium models,” and others have used vector autoregressive models for the same purpose.⁵ There has also been a significant literature on conditions under which large fiscal contractions can be expansionary.⁶ However, when compared with recent work on monetary policy, the volume of analysis is small.

7. Macroeconomists have developed two theoretical approaches to break Ricardian equivalence, e.g., to allow fiscal policies other than government spending-notably, lump-sum taxes and transfers-to have real effects, even though households are optimizing subject to intertemporal budget constraints. The first, which is the focus of this chapter, assumes that consumers have finite lives and, therefore, discount the future more rapidly than implied by the government’s budget constraint.⁷ As a result, a tax cut (or an increase in transfers) has an expansionary effect on consumption because the net present value of the tax cut exceeds that of the subsequent increase in taxes needed to keep the government solvent. The alternative approach, which is simpler theoretically and is also investigated in this chapter, assumes that some households have full ability to participate in financial markets and can intertemporally smooth consumption, while others are subject to credit constraints and cannot participate in any type of asset market. These households, therefore, just consume their after-tax disposable incomes in each period.⁸

Basic Model

8. To simplify the modeling, we assume that the economy is in a stationary steady state, so income does not trend over time and deaths equal births in each period. Utility is quadratic, which ensures certainty equivalence. Crucially, in addition to the usual discount rate, β (assumed equal to the real interest rate), consumers face an additional discount wedge, λ, reflecting the probability of death. The assets/liabilities of the dead are transferred outside of the model. Finally, we assume that income follows a first order autoregressive process.⁹

9. The consumer’s problem is

\begin{array}{l} M a x \sum_{i = 0}^{\infty} \frac{U (c_{t + i})}{(r + λ)} \\ \begin{matrix} s . t . \sum_{i = 0}^{\infty} \frac{c_{t + 1}}{(1 + r)} = \sum_{i = 0}^{\infty} \frac{y_{t + 1}}{(1 + r)} & (1.1) \end{matrix} \\ Δ y_{t + i} = θ^{y} y_{t + i - 1} + ε_{t + i}^{y} \\ U (c_{t + i}) = c_{t + i} - Γ c_{t + i}^{2} \end{array}

where y is income, c is consumption, r is the real interest rate prevailing on the market, ε^y is an unexpected shock to income, Δ is the first difference operator, and Greek letters reflect underlying parameters. Note that in this model the probability of death is equal to $\frac{λ}{(1 + λ)},$

where λ is an unknown parameter to be estimated.

10. The resulting path for consumption depends on whether the individuals were “alive” last period or not. If they were “alive” then the following equation applies:

\begin{matrix} Δ c_{t} = \frac{λ}{r + λ - θ^{y}} Δ y_{t} + \frac{r}{r + λ - θ^{y}} ε_{t}^{y} & (1.2) \end{matrix}

11. If they are “born” this period, the equation is similar except it does not include lagged values:

\begin{matrix} c_{t} = \frac{r + λ}{r + λ - θ^{y}} y_{t} & (1.3) \end{matrix}

12. Weighting the two equations appropriately implies the following aggregate consumption function:

\begin{array}{r} Δ c_{t} = \frac{λ}{r + λ - θ^{y}} Δ y_{t} - \frac{λ}{1 + λ} (c_{t - 1} - \frac{λ}{r + λ - θ^{y}} y_{t - 1}) + \frac{r}{r + λ - θ^{y}} ε_{t}^{y} - \frac{λ}{1 + λ} \frac{r}{r + λ - θ^{y}} (y_{t} - ε_{t}^{y}) \\ (1.4) \end{array}

13. The change in consumption depends on: (i) the change in income, reflecting the excess discount rate; (ii) an error correction mechanism, because of the “birth” of new individuals, whose level of consumption is heavily determined by current income; (iii) the error on income familiar from the random walk model of consumption;¹⁰ and (iv) a second order term that reflects the difference in saving behavior between those who were and were not born that period.

Fiscal Policy

14. The crucial difference between fiscal policy and changes in income is that the government’s budget constraint needs to be satisfied. Hence, a cut in taxes (net of transfers) that boosts income will need, at some point, to be counterbalanced by a future increase in taxes. We model this by assuming that, like incomes, taxes follow a first order autoregressive process, but that the trajectory is relative to a long-term level of taxes, t*, that reflects the additional cost of unanticipated changes in net taxes. Specifically,

\begin{matrix} Δ (t_{t} - t_{t}^{*}) = - θ^{τ} (t_{t - 1} - t_{t - 1}^{*}) + ε_{t}^{τ} & (1.5) \end{matrix}

Where $t_{t}^{*} = t_{t - 1}^{*} - \frac{r}{r + θ^{τ}} ε_{t}^{τ} .$

15. Hence, an unexpected fall in taxes is simultaneously accompanied by a permanent increase in the expected permanent tax rate from this point forward. The resulting consumption function looks very much like the earlier one except that unanticipated cuts in taxes $(ε_{t}^{τ})$ lower consumption through a Ricardian offset, whereas unexpected increases in income $(ε_{t}^{y})$ raise consumption through higher saving:

\begin{matrix} Δ c_{t} = \frac{λ}{r + λ - θ^{y}} Δ y_{t} - \frac{λ}{1 + λ} (c_{t - 1} - \frac{λ}{r + λ - θ^{y}} y_{t - 1}) + \frac{r}{r + λ - θ^{y}} ε_{t}^{y} - \frac{λ}{1 + λ} \frac{r}{r + λ - θ^{y}} (y_{t} - ε_{t}^{y}) \\ \begin{matrix} - \frac{λ}{r + λ - θ^{t}} Δ t_{t} - \frac{λ}{1 + λ} (c_{t - 1} - \frac{λ + r}{r + λ - θ^{t}} t_{t - 1}) + \frac{λ r}{(r - θ^{τ}) (r + λ - θ^{y})} ε_{t}^{t} & (1.6) \end{matrix} \end{matrix}

Credit Constrained Consumers

16. An alternative theoretical approach allowing fiscal policies to have real effects is to assume that all consumers are infinitely lived, but a proportion of them, η, have no access to credit markets and can thus just consume their after-tax disposable income. Again, there are two consumption processes. The credit unconstrained consumers follow the random walk model:

\begin{matrix} Δ c_{t} = \frac{r}{r + θ^{y}} ε_{t}^{y} & (1.7) \end{matrix}

while constrained individuals consume all of their disposable income:

\begin{matrix} c_{t} = y_{t} . & (1.8) \end{matrix}

17. This results in the following aggregate consumption function:

\begin{matrix} Δ c_{t} = η Δ y - η (c_{t - 1} - y_{t - 1}) + (1 - η) \frac{r}{r + θ^{y}} ε_{t}^{y} . & (1.9) \end{matrix}

18. One can also add credit constrained consumers to the myopic model discussed above. The resulting equation is obtained by simply substituting the consumption process in equation (1.6) for the random walk model in the equation (1.9) above.

Supply effects

19. It is often argued that, in addition to their direct effects on intertemporal consumption choices, increases in taxes (or cuts in transfers) have negative supply effects coming from disincentives to work. This is relatively easy to model in our framework, as t* reflects this long-term change in the burden coming from government. Assuming that long-term income falls by some proportion, γ, of the implied permanent level of taxes, this adds a further term in the unexpected change in taxes to the consumption function described by equation (1.6):

\begin{matrix} - \frac{γ r}{r + θ^{t}} ε_{t}^{t} & (1.10) \end{matrix}

C. Some Estimates

20. Empirically, an unrestricted version of the consumption model derived in the last section is estimated first. The coefficient restrictions implied by the presence of myopic and credit constrained consumers are then tested. Specifically, the deep parameters are estimated for the myopic model, and various additional considerations such as including credit constraints and supply effects are subsequently explored.

21. The model was estimated from 1960 using annual data for real consumption of nondurable goods and services, personal income excluding transfers, payments of direct taxes less transfers, disposable income (income minus net taxes). Annual data were used because taxes are levied on yearly income and it simplifies the time series characterization of the data (Figure 1), while official ISTAT data have been extended backwards to 1960 using OECD household sector data. Estimates take into account the presence of a break in the growth rate of real personal income in 1976, by allowing for parameter shifts in both the income and the tax equations.¹¹

22 The estimated unrestricted system comprises:

\begin{array}{l} Δ c_{t} = α^{c} + β^{y} Δ y_{t} + β^{t} Δ t_{t} + β^{e c m} (c_{t - 1} - μ (y_{t - 1} + t_{t - 1})) + ε_{t}^{c} \\ \begin{matrix} Δ y_{t} = α^{y} + γ^{t r e n d} t r e n d + γ^{e c m} y_{t - 1} + ε_{t}^{y} & (1.11) \end{matrix} \\ Δ t_{t} = α^{t} + τ^{y} Δ y_{t} + τ^{e c m} t_{t - 1} + ε_{t}^{t} \end{array}

where c and y are the logarithm of consumption and income while t is the net tax rate (net taxes as a ratio to income).

23. These equations correspond to the theoretical specification derived above taking account of the following considerations. In the net tax rate equation, because the tax and transfer system is progressive and the rate varies over the cycle, the growth of income is included. Also, the endogenous evolution in the equilibrium tax rate, which depends on the dynamic path of net taxes over time, is ignored. In the income equation, a time trend is included to take account of the steady rise in income over time, so the autoregressive process refers to deviations from this trend. In the consumption function, the terms in the unexpected innovation in income/net taxes are assigned to the coefficient on the change in income/taxes as attempts to estimate these terms separately produced extremely high standard errors due to collinearity. In addition, as the upward trend in both income and consumption implies that their levels are nonstationary, the coefficient on disposable income in the error correction mechanism is set at unity—the value implied by a nonstationary income process. Finally, second order terms associated with the saving of the newly born are dropped.

24. Results from estimating the unrestricted model are reported in Table 1. To test the robustness of the results, the model was estimated both directly (using seemingly unrelated regressions, hereafter SUR) and instrumental variable techniques (using the Generalized Method of Moments, or GMM). Instrumental variables are often used in consumption regressions in order to eliminate the impact of unexpected innovations in income on the specification.¹² Estimating the model with and without this effect provides a useful check on the empirical plausibility of the model. The instruments comprise all of the independent variables except the contemporaneous change in income, which was substituted by its first two lags.

Table 1.

Italy: Estimates of Unrestricted Model

Notes: Instrumental variable estimates used system GMM with instruments compraising all independent variables except the change in income, plus the first two lags of this change. Non-instrumental variable model was estimated using seemingly unrelated regressions. One and two asterisks denote that the coefficient is different from zero at 5 and 1 percent significance level, respectively.

25. Table 1 reports the coefficient estimates and fit of the unrestricted system estimated using both techniques. The SUR results reported in the first column imply that consumers spend less than one-tenth of the change in their income, while there is no significant change in consumption following a change in net taxes.¹³ It also implies that any deviation between the underlying level of consumption and disposable income is reversed at a rate of about 16 percent a year. The equation for income implies that any unexpected disturbances revert to trend at a rate of around 35 percent a year—implying a half life of less than two years. Interestingly, in the net tax rate equation, revenues fall by about 4 cents on a euro increase in income—indicating that the personal tax and transfer system is in fact not progressive—while underlying changes in the net tax rate are extremely short-lived, reverting to trend at a rate of about 35–40 percent a year. The consumption and income equations fit relatively well, with R-squares of 0.66 and 0.85, respectively, and little evidence of correlation in the residuals. The equation for net taxes fares much worse, with changes in personal income and convergence dynamics being able to explain only a limited share of the total variation in (net) tax rates.

26. The GMM results in the second column are generally similar. As expected, the impact of change in income on taxes rises somewhat, but standard errors remain too large for it to be significant. Other coefficients are essentially unchanged, with a one euro change in income resulting in less than 10 cents impact on consumption.

27. Wald tests indicate that—if tested individually—neither the hypothesis of myopic consumers nor the assumption of credit constrained consumers can be rejected by the data at conventional levels, using either SUR or GMM. Wald tests of the coefficient restriction implied by the two models are reported in Table 1 (assuming that the real interest rate is 4 percent a year). As discussed earlier, given the high collinearity of the two series, the impact of unexpected disturbances to income and net taxes are included in the coefficients on changes in these terms. Hence, the restrictions implied by the myopic model (with those included in the SUR estimates but not in the GMM ones in brackets) are:

\begin{array}{l} β^{y} = \frac{λ (+ r)}{r + λ - y^{e c m}} \\ \begin{matrix} β^{t} = \frac{λ}{λ - τ^{e c m}} (1 - \frac{r}{1 + r + τ^{e c m}}) & (1.12) \end{matrix} \\ β^{e c m} = \frac{λ}{1 + λ} \end{array}

28. Table 2 reports results from estimating the deep parameter of the myopic model-the wedge on the discount rate-using SUR and GMM. The specification for consumption, which explicitly includes innovations to income and net taxes, is as follows (in the GMM results, the coefficient on $ε_{t}^{y}$ is excluded):

\begin{matrix} Δ c_{t} = α^{y} + \frac{λ}{r + λ - y^{e c m}} Δ y_{t} (+ \frac{r}{r + λ - y^{e c m}}) + \frac{λ}{r + λ - τ^{e c m}} Δ t_{t} - \frac{r}{r + τ^{e c m}} ε_{t}^{t} + ε_{t}^{c} & (1.13) \end{matrix}

To compare these results with the unrestricted coefficient estimates reported in Table 1, the implied coefficients on the change in income $(β_{t}^{y})$ , change in net taxes $(β_{t}^{τ})$ , and error correction mechanism $(β_{t}^{e c m})$ are reported.

Table 2.

Italy: Estimates of Restricted Model with Myopic Consumers

Notes: See Table 1.

29. Both SUR and GMM results imply a statistically significant excess discount rate of 6 percent for the private sector, hence rejecting the fully-Ricardian model. The implied discount rate for Italian consumers seems consistent with the range of discount rates considered by the vast literature on retirement choices.¹⁴ At the same time, corresponding estimates for the United States are found to be much higher, suggesting that Italian consumers are prone to be much more patient than their US counterparts.

30. Changes in net taxes are found to be somewhat more persistent in the restricted model than the unrestricted model (the rate of convergence falls from 45 to 30 percent a year) while the dynamics of the tax rate are essentially unaffected. The implied coefficients for the unrestricted regressions are all extremely close to the freely estimated values, consistent with the results from the Wald test, and the fit of the model is largely unaffected. The implied SUR coefficients on the chan Table 1, but remain within one-and-a-half standard deviations of the unrestricted values in all cases.

31. Further, the myopic model is extended to allow for a proportion θ of consumers to be credit constrained. The results, reported in Table 4, suggest that including credit constraints provides no benefit to the myopic model. In this specification, both the proportion of credit constrained consumers and the excess discount rate are poorly estimated, so that the model becomes observationally equivalent to a purely Ricardian one.

Table 3.

Italy: Estimates of Restricted Model with Credit Constrained Consumers

Notes: See Table 1.

Table 4.

Italy: Estimates of Restricted Model with Myopic Consumers and Supply Effects

Notes: See Table 1. β^s was estimated using grid search methods.

32. The possible role of supply effects is also investigated. This is done by adding a negative supply effect in addition to the losses to consumption from a higher long-term net tax rate. To simplify interpretation of the coefficient on the supply effect, it is calculated as a multiple of this permanent change in taxes. Hence, for example, a coefficient of ½ implies that supply effects lower consumption by half of the long-term increase in taxes.

33. Unfortunately, it proved impossible to estimate the supply terms directly due to simultaneity of the regressors. Instead, a grid search is used to identify the coefficients that minimized the Wald test of coefficient restrictions on the unrestricted model. This procedure implies extremely small coefficients of 0.05 for the direct estimation and 0.10 for the instrumented regression. Both estimates are well below the magnitude of the parameter on the disincentive to work from taxes assumed in many microfounded “dynamic stochastic general equilibrium” models relying on the Frisch elasticity of labor supply.¹⁵

34. As can be seen in Table 4, imposing these coefficients results in extremely similar estimates of the other parameters in the model. Consequently, the only real change in the properties of the equation is that the implied coefficient on the change in net taxes is reduced in absolute value, from -0.11 to -0.07 in the direct estimation and -0.13 to -0.09 when instruments are used. In sum, while plausible supply effects are statistically indistinguishable from the basic model, they do somewhat reduce the implied size of fiscal multipliers.

D. Some Analysis

35. A fundamental feature of the intertemporal model used in this chapter is that the impact of a change in income/net taxes on consumption depends on its characteristics—its persistence, whether it is anticipated or not, and the degree to which consumers are myopic. This section explores these interactions in more detail.

36. First, the impact of a change in income/net taxes is assessed for any given level of myopia. In particular, the wedge of the discount rate over the real interest rate is assumed to be 6 percent, in line with the parameter values reported in Table 2. Figure 2 graphs how the impact of a change in disposable income on consumption varies with its type (underlying income or net taxes), its persistence (measured on the x-axis), and whether it is anticipated or not. The upper line shows the effect of an unanticipated change in underlying income, which rises steadily from around 10 cents per euro for a temporary change to a one-for-one impact if the change is permanent, with rates of convergence of income disturbances of 50, 25, and 10 percent a year giving rise to consumption multipliers of one-sixth, one-fourth, and one-half, respectively. The impact of a fully anticipated change in income (or net taxes) follows a similar path, but the effects are shrunk by about one-half as there is no boost to consumption from unanticipated saving.

37. The effect of an unanticipated change in net taxes is even lower because of the Ricardian offset. The net tax multiplier rises from around 5 cents per euro to peak at just below 20 cents for a shock that converges at 10 percent a year. At convergence rates below 5 percent, the multiplier starts to fall as the Ricardian offset increases more rapidly. Indeed, it falls to zero for a “permanent” shock to net taxes, as this violates the intertemporal budget constraint and hence the “change” in taxes is fully offset by the movement in the long-term tax rate. The difference in consumption multipliers coming from unanticipated increases in underlying income and from equivalent changes in net taxes rises steadily as the changes become more persistent—from 4 cents for a temporary disturbance to 7, 14, and 29 cents at convergence rates of 50, 25, and 10 percent, respectively—as the Ricardian offset becomes more pertinent.

38. The impact of an unanticipated change in income varies with its own persistence as well as with the level of consumers’ myopia. The upper panel in Figure 3 plots the size of fiscal multipliers associated with an unanticipated change in income, as the private sector discount wedge is varied from 2 to 12 percent—the span of estimates in the range of ± two standard errors. The impact of unanticipated changes in income rises as the level of myopia increases, and, even though the multipliers all converge to unity for a permanent change in income, these differences are quite persistent across plausible levels of income persistence. For example, the difference in income multipliers implied by a 6 and 12 percent wedge rises slowly from 5 cents per euro for a temporary income disturbance to around 12 cents for disturbances with moderate to long levels of persistence (from 40 percent a year to 10 percent). The difference falls rapidly only at persistence levels lower than 5 percent a year, but the longevity of such processes appears implausible (the half life of a change is well over a decade).

39. The impact of net taxes on consumption also increases with myopia. The lower panel of Figure 3 repeats this exercise for unanticipated changes in net tax rates rather than income. Temporary tax changes raise consumption by 2–10 cents per euro depending on the size of the discount wedge, and these changes peak at 10–33 cents for convergence rates slightly below 10 percent. Again, the differences in multipliers produced by different values of the wedge in the discount rate are relatively persistent—the difference between a 6 percent and 12 percent wedge is 5–10 cents per euro for all reasonable rates of convergence.

40. The model also allows for a calculation of the dynamic effects on consumption of a policy change. As can be seen in the top panel of Figure 4, a long-lived reduction in net taxes produces an initial boost to consumption that erodes slowly before leading to a significant permanent reduction in consumption, reflecting the substantial increase in net taxes needed to pay for the implied rise in debt. By contrast, a short-lived increase in net taxes leads to a smaller boost to consumption that dissipates much faster, but the long-term effects are comparable. The middle panel shows that adding supply effects with a coefficient of 0.05 lowers the short-term benefits to consumption and raises the long-term losses. As these effects are larger for longer-lived change in net taxes, this also reduces the difference in multipliers between short- and long-term tax changes. Given the estimated low persistence of the changes in net taxes, unanticipated and anticipated shocks to taxes and transfers end up having similarly small effects on consumption, as shown in the bottom panel.

41. The final issue discussed in this chapter refers to the possibility that large fiscal contractions could be expansionary. It has been argued that fiscal consolidations can be expansionary due to their effects on private sector expectations concerning future taxation. If forward-looking consumers anticipate long-run tax reductions because of cuts in expenditure today, then they may increase expenditure now and so offset the direct effects of the fiscal contraction. This perverse effect, which has spawned a significant literature, is often ascribed to beneficial supply-side effects, such as incentives for labor participation. Our own calculations suggest that even if the large deficit were assumed relatively long-lived, the supply benefits on income would need to be several times the implied long-term change in net taxes, a result that strikes us an implausible. More likely, in our view, is that the economy is boosted by two further mechanisms. First, the expected value of the real interest rate may fall. Such a change would provide a direct boost to the economy and would also tend to increase the Ricardian offset, thereby cushioning consumption from the impact of fiscal consolidation. The second effect comes through expectations of the rate of future consolidation. As the effect of fiscal policy on consumption depends on the net present value of the future path of net taxes, a large fiscal consolidation conveying expectations that future consolidation will occur more slowly may offset the impact of higher taxes on the net present value of income.

E. Some Conclusions

42. The chapter provides estimates of fiscal multipliers based on an intertemporal model for nondurable consumption where households are myopic, namely they discount the future at a higher rate than the prevailing real rate of interest. In such a theoretical framework, the impact of any shock to income/net taxes on consumption depends on three characteristics—the persistence of the shock, whether it is anticipated or not, and the discount wedge, i.e., the consumers’ excess of discount with respect to the market rate.

43. The estimated excess rate of discount is of the order of 6 percent—broadly consistent with previous findings for Italy indicating negligible direct effects from changes in taxes or transfers on consumption and, thereby, on growth. The model does not produce a single estimate of the multiplier associated with (say) income taxes. Rather, this value can vary between 5 and 20 cents per euro, depending on longevity of the disturbance and the degree to which it is anticipated. Historically, shocks to the net tax rate have been extremely short-lived, implying point estimates in the low end of the range. Strikingly, adding credit constrained consumers to the myopic model does not improve the fit. On the other hand, adding supply effects generates some reductions in the estimated multipliers, but the results are statistically indistinguishable from the baseline model.

44. A low discount wedge has broad policy implications. If most people discount the future at a rate just slightly higher than the borrowing rate for the government, the direct real effects of fiscal policy on consumption are likely to be limited. Similarly, the implied low discount wedge suggests that households might smooth the future expected effects of pension reforms over their lifetime consumption by engaging in greater accumulation of assets.

45. Overall, the main advantage of our framework is that it brings the intertemporal nature of disturbances to income, taxes, and transfers back to the fore of analysis. Thinking of fiscal policy in an intertemporal setting provides a range of insights. For example, to the extent that automatic stabilizers associated with the economic cycle are less persistent than other policy changes, they will be associated to smaller multipliers. Similarly, while it is unlikely that large fiscal contractions could be expansionary purely due to supply effects, it is possible that reductions in the real interest rate and changes in the assumed longevity of future tax cuts also play a role. Last but not least, the theoretical framework employed in the paper also allows one to distinguish the short-sightedness of consumers—which is mirrored by the wedge between their own discount rate and the borrowing rate of the government—from short-sightedness of fiscal policy—which determines the risk premium on government bonds. Such a distinction is crucial both from a normative and from a positive standpoint. Indeed, while the former can be seen as a structural parameter reflecting preferences, the latter is policy-determined and is therefore unlikely to be time invariant.

References

Banca d’Italia, 1976, Relazione Annuale: Assemblea Generale Ordinaria dei Partecipanti (Rome: Bank of Italy).
- Search Google Scholar
- Export Citation
Barro, Robert, 1974, “Are Government Bonds Net Wealth?” Journal of Political Economy, Vol. 82, pp. 1095 –118.
- Search Google Scholar
- Export Citation
Bayoumi, Tamim, 2004, “GEM: A New International Macroeconomic Model”, IMF Occasional Paper No. 239 (Washington DC: International Monetary Fund).
- Search Google Scholar
- Export Citation
Blanchard, Olivier, 1985, “Debt, Deficit, and Finite Horizons,” The Journal of Political Economy, Vol. 93, No. 2, pp. 223 –47.
- Search Google Scholar
- Export Citation
Blanchard, Olivier, and Roberto Perotti, 2002, “An Empirical Characterization of the Dynamic Effects of Changes in Government Spending and Taxes on Output,” The Quarterly Journal of Economics, Vol. 117, No. 4, pp. 1329 –1368.
- Search Google Scholar
- Export Citation
Botman, Dennis, Douglas Laxton, Dirk Muir, and Andrei Romanov, 2005, “Global Fiscal Model (GFM),” (unpublished; Washington DC: International Monetary Fund).
- Search Google Scholar
- Export Citation
Bryant, Ralph, Dale Henderson, Gerald Holtham, Peter Hooper, and Steven Symansky, (eds.), 1988, Empirical Macroeconomics for Interdependent Economies (Washington DC: Brookings Institution).
- Search Google Scholar
- Export Citation
Campbell, John, and Gregory Mankiw, 1990, “Permanent Income, Current Income, and Consumption,” Journal of Business and Economic Statistics, Vol. 8, pp. 265 –280.
- Search Google Scholar
- Export Citation
Coenen, Günter, and Roland Straub, 2005, “Does Government Spending Crowd In Private Consumption? Theory and Empirical Evidence for the Euro Area”, IMF Working Paper No. 05/159 (Washington DC: International Monetary Fund).
- Search Google Scholar
- Export Citation
Coenen, Günter, Peter McAdam, and Roland Straub, 2005, “Tax Reform and Labour-Market Performance in the Euro Area: A Simulation-Based Analysis Using the New Area-Wide Model” (unpublished, Frankfurt: European Central Bank).
- Search Google Scholar
- Export Citation
Erceg, Christopher, Luca Guerrieri, and Christopher Gust, 2005, “SIGMA: A New Open Economy Model for Policy Analysis,” International Finance Discussion Papers No. 835, Board of Governors of the Federal Reserve System (Washington DC: Federal Reserve Board)
- Search Google Scholar
- Export Citation
Fatas, Antonio, and Ilian Mihov, 2001, “Government size and automatic stabilizers: international and intranational evidence,” Journal of International Economics, Vol. 55, No. 1, pp. 3 –28.
- Search Google Scholar
- Export Citation
Galí, Jordi, David López-Salido, and Javier Vallés Liberal, 2005, “Understanding the Effects of Government Spending on Consumption”, CEPR Discussion Paper No. 5212 (London, UK: Centre for Economic and Policy Research).
- Search Google Scholar
- Export Citation
Ganelli, Giovanni, 2005, “The new open economy macroeconomics of government debt”, Journal of International Economics, Vol. 65, No. 1, pp. 167 –84.
- Search Google Scholar
- Export Citation
Giavazzi, Francesco, Tullio Jappelli, Marco Pagano, and Marina Benedetti, 2005, “Searching for Non-Monotonic Effects of Fiscal Policy: New Evidence”, CEPR Discussion Paper No. 5272 (London, UK: Centre for Economic and Policy Research).
- Search Google Scholar
- Export Citation
Giordano, Raffaella, Sandro Momigliano, Stefano Neri, and Roberto Perotti, 2005, “The Effects on the Economy of Shocks to Different Government Expenditure Items: Estimates with a SVAR Model” (unpublished, Rome: Bank of Italy).
- Search Google Scholar
- Export Citation
Henry, Jerome, Pablo Hernandez de Cos, and Sandro Momigliano, 2004, “The Short-Term Impact of Government Budgets on Prices—Evidence with Macroeconometric Models,” ECB Working Paper Series No. 396 (Frankfurt: European Central Bank).
- Search Google Scholar
- Export Citation
Hubbard, R. Glenn, Jonathan S. Skinner, and Stephen P. Zeldes, 1995, “Precautionary Saving and Social Insurance,” Journal of Political Economy, Vol. 103, No. 2, pp. 360 –99.
- Search Google Scholar
- Export Citation
Laxton, Douglas, and Paolo Pesenti, 2003, “Monetary Rules for Small, Open, Emerging Economies,” Journal of Monetary Economics, Vol. 50, No. 5, pp. 1109 –46.
- Search Google Scholar
- Export Citation
Leimer, Dean R., and David H. Richardson, 1992, “Social Security, Uncertainty Adjustments, and the Consumption Decision,” Economica, Vol. 59, pp. 311 –35.
- Search Google Scholar
- Export Citation
Mountford, Andrew, and Harald Uhlig, 2002, “What are the Effects of Fiscal Policy Shocks?” CEPR Discussion Paper No. 3338 (London, UK: Centre for Economic and Policy Research).
- Search Google Scholar
- Export Citation
Perotti, Roberto, 2005, “Estimating the Effects of Fiscal Policy in OECD Countries”, CEPR Discussion Paper No. 4842 (London, UK: Centre for Economic and Policy Research).
- Search Google Scholar
- Export Citation
Ricciuti, Roberto, 2003, “Assessing Ricardian Equivalence”, Journal of Economic Surveys, Vol. 17, No.1, pp. 55 –76.
- Search Google Scholar
- Export Citation
Samwick, Andrew, 1998, “Discount Rate Heterogeneity and Social Security Reform,” Journal of Development Economics, Vol. 57, pp. 117 –46.
- Search Google Scholar
- Export Citation
Smets, Frank and Raf Wouters, 2003, “An Estimated Dynamic Stochastic General Equilibrium Model of the Euro Area,” Journal of the European Economic Association, Vol. 1, pp. 1124 –1175
- Search Google Scholar
- Export Citation

Prepared by Silvia Sgherri (EUR).

See, among others, Ricciuti (2003) for a review, the seminal work by Barro (1974), and the contribution by Campbell and Mankiw (1990).

Clearly, changes in the tax wedge are also likely to affect aggregate demand through their effects on labor utilization. Such an effect will depend on the particular characteristics of the labor market at hand, e.g., the elasticity of labor supply and labor demand and the details of the wage-setting process (Coenen and others, 2005).

See, for example, the recent study by Giordano and others (2005) and the evidence from macroeconometric models summarized in Henry and others (2004).

Bryant and others (1988) includes results from a range of traditional macroeconomic models. Vector autoregression analysis includes Blanchard and Perotti (2002), Fatás and Mihov (2001), Moutford and Uhlig (2002), and Perotti (2002). A summary of results is contained in IMF (2003). Bayoumi (2004) discusses the new approach to large macroeconomic models embodied in “stochastic dynamic equilibrium models”, while Laxton and Pesenti (2003), Erceg and others (2004), and Smets and Wouters (2004) describe such models. The IMF’s Global Fiscal Model, described in Ganelli (2004) and Botman and others (2005), is the only one of the new generation of models primarily designed for fiscal policy analysis.

Giavazzi and others (2005) provide a recent review of empirical studies on expansionary fiscal contractions.

It should be stressed that the term “death” or “finite life” defines economic death rather than its physical counterpart. This can occur through unexpected events that make previous optimal consumption plans irrelevant—examples would include winning the lottery, or a sudden and unexpected job loss or bankruptcy.

Galé and others (2005) and Coenen and Straub (2005) have extended the standard New-Keynesian sticky-price model by allowing for the coexistence of “non-Ricardian” and “Ricardian” households.

The model can be easily generalized to other income processes and assumptions. The current framework is utilized as it provides a simple closed form estimating equation.

Indeed, if A is set equal to zero, the model simplifies to a random walk.

Following the standardization of the wage indexation system in conjunction with sizeable wage increases upon renewals of contracts in the industry sector, average real personal income surged dramatically in 1976. Personal income taxes increased accordingly, implying no change in the average tax rate and disposable income growth. See Banca d’Italia (1976).

We assume underlying tax rates are known ahead of time, although the short-term impact of changes in income on tax rates is instrumented.

The (unrestricted) response of consumption to changes in taxes and transfers remains insignificant even when a system allowing for separated equations for taxes and transfers is considered.

See, among others, Samwick, 1998; Hubbard and others, 1996; Leimer and Richardson, 1992.

Laxton and Pesenti (2003), for example, consider a Frisch elasticity of labor supply equal to 0.3, whereas Coenen and others (2005) calibrate the parameter at 0.5.

Other Publishers

Save
Cite
Email this content

Share Link

Copy this link, or click below to email it to a friend
Email this content
or copy the link directly:

https://www.elibrary.imf.org/view/journals/002/2006/059/article-A004-en.xml
The link was not copied. Your current browser may not support copying via this button.

Link copied successfully

Italy: Selected Issues

Author: International Monetary Fund

Volume/Issue:: Volume 2006: Issue 059

Publisher:: International Monetary Fund

ISBN:: 9781451819885

ISSN:: 1934-7685

Pages:: 142

DOI:: https://doi.org/10.5089/9781451819885.002

Keywords
IV. How Expansionary Are Tax Cuts in Italy?1
References

View raw image

Italy: The Data

View raw image

Italy: Impact of Changes in Disposable Income

View raw image

Italy: Impact of Changes in Disposable Income for Different Degrees of Myopia

View raw image

Italy: Impact of Lower Net Taxes

$\begin{array}{l} Δ c_{t} = α^{c} + β^{y} Δ y_{t} + β^{t} Δ t_{t} + β^{e c m} (c_{t - 1} - μ (y_{t - 1} + t_{t - 1})) + ε_{t}^{c} \\ Δ y_{t} = α^{y} + γ^{t r e n d} t r e n d + γ^{e c m} y_{t - 1} + ε_{t}^{y} \\ Δ t_{t} = α^{t} + τ^{y} Δ y_{t} + τ^{e c m} t_{t - 1} + ε_{t}^{t} \end{array}$
	No Instrumental	Instrumental
	Variables	Variables
Consumption equation
α^c	.48 (.08)^**	.49 (.01)^**
μ	.40 (.13) ^**	.42 (.09) ^**
β^y	.08 (.04) ^**	.08 (.08) ^**
β^t	-.38 (.35)	-.48 (.38)
β^ecm	-.16 (.07) ^**	-.17 (.06)^**
R²	.66	.66
DW	1.45	1.41
Income equation
α^y	1.67 (.14) ^**	1.63 (.14) ^**
γ^trend	.020 (.002) ^**	.020 (.002) ^**
γ^ecm	-.35 (.03) ^**	-.34 (.03) ^**
R²	.85	.85
DW	1.71	1.70
Net tax rate equation
α^t	-.03 (.01)^**	-.03 (.01)^**
τ^y	-.04 (.02) ^**	-.07 (.03) ^**
τ^ecm	-.45 (.12) ^**	-.41 (.12) ^**
R²	.19	.08
DW	2.11	1.96
*Wald test of coefficient restrictions*
Myopic model: χ²(2)	4.0	5.2
Credit constrained consumers: χ²(2)	5.9	5.5

$\begin{array}{l} Δ c_{t} = α^{c} + β^{y} Δ y_{t} + β^{t} Δ t_{t} + β^{e c m} (c_{t - 1} - μ (y_{t - 1} + t_{t - 1})) + ε_{t}^{c} \\ Δ y_{t} = α^{y} + γ^{t r e n d} t r e n d + γ^{e c m} y_{t - 1} + ε_{t}^{y} \\ Δ t_{t} = α^{t} + τ^{y} Δ y_{t} + τ^{e c m} t_{t - 1} + ε_{t}^{t} \end{array}$
	No Instrumental	Instrumental
	Variables	Variables
Consumption equation
α^c	.48 (.08)^**	.49 (.01)^**
μ	.40 (.13) ^**	.42 (.09) ^**
β^y	.08 (.04) ^**	.08 (.08) ^**
β^t	-.38 (.35)	-.48 (.38)
β^ecm	-.16 (.07) ^**	-.17 (.06)^**
R²	.66	.66
DW	1.45	1.41
Income equation
α^y	1.67 (.14) ^**	1.63 (.14) ^**
γ^trend	.020 (.002) ^**	.020 (.002) ^**
γ^ecm	-.35 (.03) ^**	-.34 (.03) ^**
R²	.85	.85
DW	1.71	1.70
Net tax rate equation
α^t	-.03 (.01)^**	-.03 (.01)^**
τ^y	-.04 (.02) ^**	-.07 (.03) ^**
τ^ecm	-.45 (.12) ^**	-.41 (.12) ^**
R²	.19	.08
DW	2.11	1.96
*Wald test of coefficient restrictions*
Myopic model: χ²(2)	4.0	5.2
Credit constrained consumers: χ²(2)	5.9	5.5

$\begin{matrix} Δ c_{t} = α^{c} + \frac{1}{r + λ - γ^{e c m}} (λ Δ y_{t} + r ε_{t}^{y}) - \frac{λ}{r + λ - τ^{e c m}} (1 - \frac{r}{r - τ^{e c m}}) Δ t_{t} \\ - \frac{λ}{1 + λ} (c_{t - 1} - (y_{t - 1} + t_{t - 1})) + ε_{t}^{c} \\ Δ y_{t} = α^{y} + γ^{t r e n d} t r e n d + γ^{e c m} y_{t - 1} + ε_{t}^{y} \\ Δ y_{t} = α^{y} + γ^{y} Δ y_{t} + τ^{e c m} t_{t - 1} + ε_{t}^{t} \end{matrix}$
	No Instrumental	Instrumental
	Variables	Variables
Consumption equation
α^c	.37 (.04) **	.04 (.01) **
λ	.06 (.00) **	.06 (.02) **
γ^ecm	-.36 (.03) **	-.34 (.03) **
τ^ecm	-.30 (.08) **	-.38 (.14) **
R²	.64	.66
DW	1.65	1.55
Income equation
α^y	1.74 (.15) **	1.66 (.15) **
γ^trend	.016 (.002) **	.016 (.002) **
R²	.80	.81
DW	1.48	1.47
Net tax rate equation
α^t	-.02 (.01)**	-.03 (.01)**
τ^y	-.04 (.02) **	-.06 (.04) *
R²	.18	.12
DW	2.13	2.03
*Implied Coefficients*
̂β^y	.22	.13
̂β^t	-.13	-.11
̂β^ecm	-.06	-.06

$\begin{matrix} Δ c_{t} = α^{c} + \frac{1}{r + λ - γ^{e c m}} (λ Δ y_{t} + r ε_{t}^{y}) - \frac{λ}{r + λ - τ^{e c m}} (1 - \frac{r}{r - τ^{e c m}}) Δ t_{t} \\ - \frac{λ}{1 + λ} (c_{t - 1} - (y_{t - 1} + t_{t - 1})) + ε_{t}^{c} \\ Δ y_{t} = α^{y} + γ^{t r e n d} t r e n d + γ^{e c m} y_{t - 1} + ε_{t}^{y} \\ Δ y_{t} = α^{y} + γ^{y} Δ y_{t} + τ^{e c m} t_{t - 1} + ε_{t}^{t} \end{matrix}$
	No Instrumental	Instrumental
	Variables	Variables
Consumption equation
α^c	.37 (.04) **	.04 (.01) **
λ	.06 (.00) **	.06 (.02) **
γ^ecm	-.36 (.03) **	-.34 (.03) **
τ^ecm	-.30 (.08) **	-.38 (.14) **
R²	.64	.66
DW	1.65	1.55
Income equation
α^y	1.74 (.15) **	1.66 (.15) **
γ^trend	.016 (.002) **	.016 (.002) **
R²	.80	.81
DW	1.48	1.47
Net tax rate equation
α^t	-.02 (.01)**	-.03 (.01)**
τ^y	-.04 (.02) **	-.06 (.04) *
R²	.18	.12
DW	2.13	2.03
*Implied Coefficients*
̂β^y	.22	.13
̂β^t	-.13	-.11
̂β^ecm	-.06	-.06

$\begin{matrix} Δ c_{t} = α^{c} + [θ + \frac{(1 - θ) λ}{r + λ - γ^{e c m}}] Δ y_{t} + \frac{(1 - θ) λ}{r + λ - γ^{e c m}} ε_{t}^{y} - {θ + (1 - θ) \frac{r}{r + λ - τ^{e c m}} [1 - \frac{r}{r - τ^{e c m}}]} Δ t_{t} \\ - [θ + (1 - θ) \frac{λ}{1 + λ}] (c_{t - 1} - (y_{t - 1} + t_{t - 1})) + ε_{t}^{c} \\ Δ y_{t} = α^{y} + γ^{t r e n d} t r e n d + γ^{e c m} y_{t - 1} + ε_{t}^{y} \\ Δ t_{t} = α^{t} + τ^{y} Δ y_{t} + τ^{e c m} t_{t - 1} + ε_{t}^{t} \end{matrix}$
	No Instrumental	Instrumental
	Variables	Variables
Consumption equation
α^c	-.14 (.01) **	.04 (.01) **
θ	.04 (.24)	.07 (.13)
λ	.06 (.10)	.03 (.05)
γ^ecm	-.35 (.03) **	-.34 (.03) **
τ^ecm	-.34 (.08) **	-.40 (.13) **
R²	.54	.50
DW	1.53	1.39
Income equation
α^y	1.68 (.16) **	1.66 (.15) **
γ^trend	.020 (.002) **	.016 (.002) **
R²	.81	.81
DW	1.46	1.47
Net tax rate equation
α^t	-.03 (.01)**	-.03 (.01)**
τ^y	-.04 (.02) **	-.06 (.03) **
R²	.17	.10
DW	2.07	2.00

Topics

Countries

Series

About

Resources

Abstract

IV. How Expansionary Are Tax Cuts in Italy?1

A. Introduction

B. Some Theory

Basic Model

Fiscal Policy

Credit Constrained Consumers

Supply effects

C. Some Estimates

D. Some Analysis

E. Some Conclusions

References

Same Series

Other IMF Content

Other Publishers

Asian Development Bank

Food and Agriculture Organization

Inter-American Development Bank

International Labour Organization

Nordic Council of Ministers

The World Bank

Share Link

Table of Contents

Headings

Figures

Metrics

Foresee

IV. How Expansionary Are Tax Cuts in Italy?¹