Thailand: Fiscal Policy and Inclusive Growth1
A semi-structural model is used to simulate the impact of fiscal structural reforms on economic growth, inflation, current account balance, and public finances. Simulation results indicate that structural reforms enabling higher infrastructure investment, stronger labor participation, and more efficient taxation can raise growth significantly and contribute to addressing domestic and external imbalances.
1. Thailand’s cyclical and structural challenges have contributed to its high current account surplus and are expected to impact growth in the short and long term. During 2013–15, growth slowed down significantly. While the export and tourism sectors have since regained strength, other sectors of the economy have yet to take part in this renewed dynamism. Consumption and investment have remained sluggish, partly due to high household debt and a slow trickle down of growth to household income. This has resulted in a high current account surplus. At the same time, rapid population aging will soon become a drag on potential growth, as it has an impact on the size of the work force and productivity growth (Figure 1).
Figure 1.Thailand: Macro Situation
2. The current environment provides an opportunity for fiscal policy to support structural reform and stimulate the economy, while safeguarding fiscal sustainability. Under currently low international interest rates, low inflation, and sluggish domestic demand, fiscal stimulus would help bring output closer to potential, steer inflation towards the target, and support external rebalancing. Moreover, fiscal policy in support of structural reform would not only boost demand in the short term, but also lead to higher potential growth by raising the capital stock and productivity, and providing better incentives for labor force participation. As expenditure pressures are likely to rise in the longer term because of population aging, revenue mobilization should also form part of the longer-term structural agenda to safeguard fiscal sustainability.
3. Staff has recommended a number of fiscal structural reforms to support the economy, several of which are already part of the authorities’ agenda. These include improving public investment management and scaling up public infrastructure investment over the next five years, possibly frontloading some of the Eastern Economic Corridor projects (Box 1 of the Staff Report); better targeting of social expenditures to address poverty and inequality, including through the appropriate implementation of the Low-Income Earners Registry and the Government Welfare Program (Box 1); a broad strategy to prepare for expenditure pressures associated with aging; gender-inclusive policies to improve female labor force participation; and a tax reform in the longer term to mobilize resources needed to address expenditure pressures from population aging.
4. This paper focuses on the impact of key structural fiscal reforms on growth and other macro variables. It simulates the impact of: (i) a public infrastructure push; (ii) labor market policies, including an increase in the pensionable age and in provision of childcare services; and (iii) a change in the composition of taxes from income taxes to value added tax (VAT) to shed light on the desired composition of additional taxes to be levied in the longer term. The specific assumptions behind each policy package and the corresponding simulation results are described below.
B. Model Features and Key Assumptions
5. The simulations are based on APDMOD, a module of the IMF’s Flexible System of Global Models.2 This is a semi-structural model of the global economy, with individual blocks for 16 Asian countries, and 8 additional regions to represent the rest of the world. The model has a rich fiscal sector with 7 possible instruments, including spending on consumption, infrastructure, and transfers, lumpsum taxation on households, and distortionary taxation on consumption, labor, and capital. Only some, “saving”, households hold debt as a source of wealth, which allows them to smooth consumption in the face of shocks or policy changes. Other “low-income” households cannot save effectively and live only from their current income. These non-Ricardian properties allow for a powerful role for fiscal policy in both the short and long terms. Monetary policy is assumed to follow an inflation targeting regime, implemented with an interest rate reaction function based on the output gap and forecasted inflation (the “standard rule”), but can also be modified to follow alternative rules.
C. Simulations and Results
6. An infrastructure push was simulated, amounting to a cumulative 5 percentage points of GDP increase in public investment, mostly frontloaded in the first three years (Figure 2, left chart). This roughly amounts to fully implementing the projects planned under the EEC.3 In combination with the infrastructure push, three versions of the interest rate reaction function (Figure 2, right chart) are simulated: (i) normal conduct of monetary policy characterized by the standard rule, where the interest rate increases as the inflation and output gaps increase (scenario 1); (ii) monetary accommodation by keeping the policy rate unchanged with respect to the baseline scenario, throughout the five years of fiscal stimulus (scenario 2); and (iii) policy rate easing for two years followed by policy rates unchanged. both with respect to the baseline scenario (scenario 3).
Figure 2.Thailand: Infrastructure Push and Monetary Stimulus
7. The results point to a significant increase in GDP and inflation, and a substantive, albeit temporary, reduction in the current account balance. The results (Figure 3) point to a total increase in output and inflation, under monetary easing, amounting to 2.3 percentage points and 1.2 percentage points, respectively, over a five-year span. Growth increases most during the first two years, when the bulk of the infrastructure scale-up takes place, and reverses thereafter as infrastructure investment winds down.4 The impact on growth and inflation is significantly higher under monetary easing than under the standard rule that tightens monetary policy in response to the fiscal stimulus. Staff would not recommend monetary easing just to compound the fiscal impact, as monetary policy should focus on the inflation objective. However, given staff’s recommendation for monetary easing in the current juncture, combining it with fiscal stimulus would help attain inflation and growth objectives. The temporary infrastructure push also reduces the current account balance, especially during years two and three.5 Finally, the infrastructure push leads to a temporary increase in the fiscal deficit and a higher public debt, especially under scenario 1, where monetary policy is tightened. The stronger nominal growth and lower interest rates under scenarios 2 and 3 significantly attenuate the impact on public debt, which increases close to 2 percentage points in these scenarios rather than 4.4 percentage points under scenario 1.
Figure 3.Thailand: Impact of Public Infrastructure Push
Labor Sector Reform
8. Two reforms that directly impact the labor market are simulated: (i) a permanent increase in the retirement age by 2 years and (ii) an increase in the provision of childcare services by 0.25 percent of GDP per year, leading to higher female labor force participation. At 55, the retirement age in Thailand is lower than the average for Asia (57), for emerging markets in Europe (61) and Latin America (62), and for advanced economies outside of Asia (64). Increasing the retirement age raises labor force participation of the cohort directly affected.6 The budget gains from a higher retirement age are assumed to be distributed among retirees in the form of higher pensions. The cost of childcare services is assumed to be borne by saving households through a progressive increase in personal income tax (PIT).7 Although women’s participation rate in Thailand (60.8 percent) is relatively high by international perspective, it still lags that of men (77.5 percent).
9. The impact of labor market reforms on real GDP can be large. Figure 4 shows the impact of these measures on real GDP over a 5-year period. The two measures together lead to a 0.6 percentage point increase in real GDP, with child-care services having a larger contribution.
Figure 4.Thailand: Impact of Labor Market Reforms on GDP
Source: IMF staff estimates.
10. Taken together, the infrastructure investment and labor market reform can have a significant impact on output in the short and long term, especially under monetary easing. The joint impact of the infrastructure scale-up and labor market reform on output over a 5-year period would amount to 3 percent (Figure 5) assuming monetary easing.
Figure 5.Thailand: Impact of Fiscal/Structural Reforms on GDP
Source: IMF staff estimates.
Note: Assumes some monetary easing.
Tax Reform: Value Added vs. Income Taxes
11. Given that revenue mobilization is likely to be necessary in the long term, an informed decision will be needed on which taxes to raise. The design of the VAT in Thailand is broadly in line with best practices, with a single rate, broad base, and low compliance gaps. However, the rate of 7 percent is one of the lowest in the world and, therefore, an increase in this rate would be an adequate option for revenue mobilization. While the VAT is generally estimated to produce a smaller drag on output than income taxes, it can be more regressive. This section looks at the additional output obtained when using the VAT instead of income taxes to raise additional fiscal revenues and the possibility of compensating low-income earners to attenuate the impact of VAT on the lower incomes deciles.8 There are two scenarios: (i) VAT collection is increased by 1.5 percent of GDP, which is offset by cuts of 0.75 percent of GDP in both PIT and corporate income tax (CIT); 9 (ii) VAT collection is increased by 1.75 percent of GDP, offset by cuts of 0.75 percent of GDP in both PIT and CIT and a 0.25 percent of GDP rebate of the VAT increase to low-income households.
12. Relying on VAT is preferable over income taxes to mobilize fiscal revenues in the future. Without compensating low-income households, a tax reform raising VAT results in a more efficient economy with 0.5 percentage points more in GDP over 5 years, compared to a tax reform relying on higher income taxes (Figure 6). When a rebate is provided to lower income deciles in order to protect their disposable income, the resulting real GDP is another 0.1 percent of GDP higher – in total, 0.6 percent of GDP higher than under an increase in income taxes. The stronger impact is because low-income households have a higher propensity to consume than saving households.
Figure 6.Thailand: Real GDP: Impact of Replacing Personal Income Tax with VAT
Source: IMF staff estimates.
13. Structural reforms enabling higher infrastructure investment, stronger labor participation, and more efficient taxation could help raise growth and address domestic and external imbalances. Scaling up public infrastructure investment and providing appropriate incentives for labor force participation would boost growth and inflation in the short term and reduce the current account surplus over the medium term. It would also lead to higher potential output over the longer term by raising the capital stock and productivity, and offsetting some of the drag from aging on the number of workers. As revenue mobilization will be necessary in the medium to long term, relying on VAT and compensating lower income deciles rather than relying on income taxes could lead to a more efficient tax composition and higher growth.
Box 1.The Low-Income Earners Registry and the Government Welfare Program
Thailand has reduced poverty at an impressive rate over the last three decades. The national poverty rate dropped from 67 percent in 1986 to 10 percent in 2014, while extreme poverty also declined sharply. Until recently, however, welfare transfers were universal or categorically targeted, leading to large inclusion and exclusion errors. The recently implemented government welfare program should help reduce inclusion and exclusion errors through better targeting.
The authorities are looking for ways to improve current social transfer programs. Social transfers in Thailand are heavily skewed towards contributory social insurance type benefits, which exclude informal workers and their family members. In recent years, the Thai authorities have taken a number of steps to improve the coverage and impact of their social assistance schemes, but these efforts have essentially focused on categorical benefits, including a universal old age allowance.
The government’s current development plan (National Economic and Social Development Plan 2017–21), includes the establishment of a registry of low income earners intended to facilitate more effective targeting of newly introduced social assistance benefits. Registration into this database took place during April-May 2017, through specialized financial institutions (Bank of Agriculture and Agricultural Cooperatives, Government Savings Bank, and Krung Thai Bank) and local governments. There are currently 14.2 million people registered, of which 11.4 million qualified for the state welfare benefits. Given that the current registry is based on self-declared information, venues to cross check this information are necessary. In this context, the small number of people filing personal income tax returns and the lack of reliable asset registries are a challenge.
The Government Welfare Program was rolled out in 2017 to address chronic poverty. This program feeds from the low-income earners registry, in order to determine who qualifies. Recent measures under this program include a proxy means tested scheme paying cash benefits and transportation subsidies. Under the recently rolled-out National e-Payment Master Plan, low-income earners with yearly income between 30,000 and 100,000 baht (roughly US$950 and US$3,000, respectively) could buy goods from ‘Blue Flag’ stores worth 200 baht per month using an electronic welfare card. Those with yearly income below 30,000 baht could buy up to 300 baht monthly with this card. Low-income earners also receive a discount for three months on cooking gas and may be eligible for bus and train fares each worth 500 baht. In 2018, the program is also providing occupational training to help low-income earners achieve and sustain better living standards and increase their income.
Improvements to the social welfare program should be part of the Government agenda to ensure adequate targeting of social expenditures. Two key areas of work for ensuring the effectiveness of the program are: (i) improving the ability to cross-check data provided to assess eligibility; and (ii) building a strategy based on a comprehensive view of central and local welfare schemes currently in place.
AndrleM.BlagraveP.EspaillatP.HonjoK.HuntB.KortelainenM.LalondeR.LaxtonD.MavroeidiE.MuirD.MursulaS. and SnuddenS.2015 “The Flexible System of Global Models—FSGM” IMF Working Paper 15/64 (Washington: International Monetary Fund).
BassaniniA. and DuvalR.2006 “Employment Patterns in OECD Countries: Reassessing the Role of Policies and Institutions” OECD Economics Department Working Paper No. 486 (Paris: OECD Publishing).
BouisR. and DuvalR.2011 “Raising Potential Growth After the Crisis: A Quantitative Assessment of the Potential Gains from Various Structural Reforms in the OECD Area and Beyond” OECD Economics Department Working Paper No. 835 (Paris: OECD Publishing).
Prepared by Dirk Muir and Manrique Saenz.
As described in Box 1 of the Staff Report, the baseline scenario assumes only partial implementation of the infrastructure scale-up under the EEC.
The temporary public investment scale-up increases output not only in the short run but also in the long run as it leads to a higher capital stock and higher potential GDP.
Since, under the baseline scenario (without these additional 5 percentage point of GDP in infrastructure investment), the current account balance is projected to be on a declining path, the above simulation implies that the current account surplus drops more steeply during the first few years but then converges to the baseline path over the longer term.
Estimates found in Bouis and Duval (2011), originally calculated for the OECD.
The impact of the provision of childcare services on labor force participation is calibrated based on estimates for OECD by Bassanini and Duval (2006). The impact of PIT on labor supply reflects Thailand-specific calibration of the model’s parameters.
The Low-Income Earners Registry currently being developed as part of the Government Welfare Program would be instrumental in implementing the VAT rebate to the lower deciles suggested below.
Revenue mobilization recommended by staff in the long run would include an increase in VAT rates, but not a cut in PIT and CIT, given the need for additional revenues to address expenditure pressures from population aging in the future. The implied increase in the VAT tax rate is roughly 2 percentage points to generate 1.5 percent of GDP in additional revenues.