33
Maldives National Macroeconomic Forecasting And Policy Simulation Models

National Macroeconomic Forecasting And Policy Simulation ... · •Using the Quartely Real GDP series (2003Q1 to 2015Q2) •Model using information on its own past values •AR (1

  • Upload
    others

  • View
    2

  • Download
    0

Embed Size (px)

Citation preview

Page 1: National Macroeconomic Forecasting And Policy Simulation ... · •Using the Quartely Real GDP series (2003Q1 to 2015Q2) •Model using information on its own past values •AR (1

MaldivesNational Macroeconomic Forecasting And Policy Simulation Models

Page 2: National Macroeconomic Forecasting And Policy Simulation ... · •Using the Quartely Real GDP series (2003Q1 to 2015Q2) •Model using information on its own past values •AR (1

Content

• Maldivian Economy• Forecasting Models

• Process• Tourist Arrivals• GDP• Revenues

Page 3: National Macroeconomic Forecasting And Policy Simulation ... · •Using the Quartely Real GDP series (2003Q1 to 2015Q2) •Model using information on its own past values •AR (1

Real Sector

• Real GDP per capita in 2015: US$ 6,088.68 (projected)

• High dependency on tourism ‐contributes about 30% to GDP

• Highly susceptible to  external shocks

• Share of fisheries sector is small, but still a key source of employment for many

• Unemployment rate (broad definition): 28% in 2010 ‐40%

‐20%

0%

20%

40%

60%

‐15%

‐10%

‐5%

0%

5%

10%

15%

20%

25%

2003

2004

2005

2006

2007

2008

2009

2010

2011

2012

2013

2014

2015

Tourism Sector Growth (Right Axis) Real GDP Growth (Left Axis)

Co-Movement of GDP and Tourism

Financial crisis

Indian Ocean tsunami

Page 4: National Macroeconomic Forecasting And Policy Simulation ... · •Using the Quartely Real GDP series (2003Q1 to 2015Q2) •Model using information on its own past values •AR (1

Real Sector ‐ Prices

• Inflation peaked in early 2012, reflecting the introduction of the GST and the devaluation of rufiyaa

• Gradually back to lower levels as these one off effects dissipated

• Food has a weight of 23.8% in the CPI basket

• Exchange rate pass through is extremely high, and international commodity price changes are reflected to a high degree

0%

2%

4%

6%

8%

10%

12%

14%

16%

18%

20%

Q1‐20

09Q2‐20

09Q3‐20

09Q4‐20

09Q1‐20

10Q2‐20

10Q3‐20

10Q4‐20

10Q1‐20

11Q2‐20

11Q3‐20

11Q4‐20

11Q1‐20

12Q2‐20

12Q3‐20

12Q4‐20

12Q1‐20

13Q2‐20

13Q3‐20

13Q4‐20

13Q1‐20

14Q2‐20

14Q3‐20

14Q4‐20

14Q1‐20

15Q2‐20

15

Inflation Inflation excluding fish

Introduction of GST

Rufiyaa devaluation

Inflation

Page 5: National Macroeconomic Forecasting And Policy Simulation ... · •Using the Quartely Real GDP series (2003Q1 to 2015Q2) •Model using information on its own past values •AR (1

Fiscal Sector

• Government expenditure consistently exceeds revenue

• Major revenue source was the import duty until 2011

• Now, GST (especially T‐GST) and the business profit tax are the main sources, and import duty rates have been reduced

• Largest share of government expenditure goes to employee remuneration

• 70% of total expenditure is recurrent expenditure

0%

2%

4%

6%

8%

10%

12%

14%

16%

18%

20%

0%

20%

40%

60%

80%

100%

120%

140%

160%

2004 2005 2006 2007 2008 2009 2010 2011 2012 2013 2014 2015

Total revenue and grants (% of GDP, left axis)

Total expenditure and net lending (% of GDP, left axis)

Deficit (% of GDP, right axis)

Revenue, Expenditure and Financing (% of GDP)

Page 6: National Macroeconomic Forecasting And Policy Simulation ... · •Using the Quartely Real GDP series (2003Q1 to 2015Q2) •Model using information on its own past values •AR (1

Fiscal Sector ‐ Debt

• Total debt has been rising sharply since 2008, mainly fuelled by rising domestic debt

• Newly introduced fiscal responsibility act stipulates that the debt‐to‐GDP ratio should be at most 60% by the end of 2016, and maintained below the threshold thereafter  ‐

 5,000

 10,000

 15,000

 20,000

 25,000

 30,000

 35,000

0%

10%

20%

30%

40%

50%

60%

70%

80%

2004 2005 2006 2007 2008 2009 2010 2011 2012 2013 2014

Total debt (right axis) Debt to GDP (left axis)

External Debt to GDP (left axis)

Mill

ions

of R

ufiy

aa

Page 7: National Macroeconomic Forecasting And Policy Simulation ... · •Using the Quartely Real GDP series (2003Q1 to 2015Q2) •Model using information on its own past values •AR (1

External Sector ‐ Trade

• Have consistently been experiencing trade deficits

• Current account deficit is expected to be at 7% of GDP in 2015

• Trade deficit is expected to be at 56% of GDP in 2015

• Services account surplus is expected to be at 72% of GDP in 2015 0

500

1000

1500

2000

2500

2005 2006 2007 2008 2009 2010 2011 2012 2013 2014 2015

      Domestic exports f.o.b.        Imports f.o.b. Trade deficit

Mill

ions

of U

SD

Page 8: National Macroeconomic Forecasting And Policy Simulation ... · •Using the Quartely Real GDP series (2003Q1 to 2015Q2) •Model using information on its own past values •AR (1

External Sector ‐ Reserves

• Gross international reserves consist of the reserve assets held by MMA

• Built up mainly through tourism sector activity (tax receipts) and the accumulation of foreign currency assets by banks

• Challenge to intervene in FOREX market due to low reserve levels

• Reserves improving in recent years

0

100

200

300

400

500

600

700

800

  ‐

  0.5

  1.0

  1.5

  2.0

  2.5

  3.0

  3.5

  4.0

  4.5

  5.0

2009 2010 2011 2012 2013 2014 Jun‐15

Gross international reserves (right axis)

Gross international reserves import coverage (left axis)

Mill

ions

of U

SD

Mon

ths

Page 9: National Macroeconomic Forecasting And Policy Simulation ... · •Using the Quartely Real GDP series (2003Q1 to 2015Q2) •Model using information on its own past values •AR (1

Monetary Sector – Credit Growth

• Rapid increase in credit to government after 2008 partly reflects the widening fiscal deficit

• Expansion of credit to private sector stalled after 2008

• Global financial crisis • Deterioration in the asset quality of banks (increase in NPLs)

• Diverted to investment in T‐bills   ‐

  2,000.0

  4,000.0

  6,000.0

  8,000.0

  10,000.0

  12,000.0

  14,000.0

  16,000.0

  18,000.0

2005 2006 2007 2008 2009 2010 2011 2012 2013 2014 Jul‐15

Credit to central government Credit to private sector

Mill

ions

of R

ufiy

aa

Page 10: National Macroeconomic Forecasting And Policy Simulation ... · •Using the Quartely Real GDP series (2003Q1 to 2015Q2) •Model using information on its own past values •AR (1

Monetary Sector – Interest Rates

• Interest rates increased dramatically for T‐bills of all maturities reflecting the widening fiscal deficit

• Tap system introduced in 2014 to curb the increasing interest burden on the government

• This did not reduce T‐bill investment, reflecting the lack of other investment opportunities in the domestic market

• T‐bill interest rates mechanically halved in November2015.

3

4

5

6

7

8

9

10

11

2007 2008 2009 2010 2011 2012 2013 2014 Nov‐15

28 days T‐bill 90 daysT‐bill 182 days T‐bill 364 days T‐bill

T-bill Interest Rates

Perc

ent

Page 11: National Macroeconomic Forecasting And Policy Simulation ... · •Using the Quartely Real GDP series (2003Q1 to 2015Q2) •Model using information on its own past values •AR (1

ForecastingProcess

Page 12: National Macroeconomic Forecasting And Policy Simulation ... · •Using the Quartely Real GDP series (2003Q1 to 2015Q2) •Model using information on its own past values •AR (1

Macro Modeling

• Budget Process• Macroeconomic Policy Coordination Committee • Medium‐term budgets• Policy approval

• Medium Term Economic Framework• Being built with the help of World Bank• Brings together isolate models• Behavioral linkages• Excel based

Page 13: National Macroeconomic Forecasting And Policy Simulation ... · •Using the Quartely Real GDP series (2003Q1 to 2015Q2) •Model using information on its own past values •AR (1

Forecasting ModelsGDP

Page 14: National Macroeconomic Forecasting And Policy Simulation ... · •Using the Quartely Real GDP series (2003Q1 to 2015Q2) •Model using information on its own past values •AR (1

Procedural Change

National Bureau of Statistics (NBS) to discontinue production of annual medium term GDP forecasts‐ from 2015 NBS to produce Quarterly National Accounts(QNA) –first release 20thOctober 

• NBS will produce actual GDP, quarterly and annually • Not the international best practice for statistical agencies  to produce annual forecasts

EDPD/MoFT to produce annual forecasts for GDP growth and Nominal GDP 

Page 15: National Macroeconomic Forecasting And Policy Simulation ... · •Using the Quartely Real GDP series (2003Q1 to 2015Q2) •Model using information on its own past values •AR (1

Forecasting model

Univariate time series model• Using the Quartely Real GDP series (2003Q1 to 2015Q2)• Model using information on its own past values• AR (1 ,4 ) SARIMA(0,1,0)4 model

Advantages• Simple to model • Often found to produce better forecasts than structural models• Incorporates information of the economic growth up to Q2 2015

Disadvantages• Retrospective  • Cannot dissect the growth impact of specific items• A‐theoretical

Page 16: National Macroeconomic Forecasting And Policy Simulation ... · •Using the Quartely Real GDP series (2003Q1 to 2015Q2) •Model using information on its own past values •AR (1

Forecast – Univariate time series model

Page 17: National Macroeconomic Forecasting And Policy Simulation ... · •Using the Quartely Real GDP series (2003Q1 to 2015Q2) •Model using information on its own past values •AR (1

GDP GROWTH:2015‐2018

4.00%

4.50%

5.00%

5.50%

6.00%

6.50%

7.00%

7.50%

8.00%

2015 2016 2017 2018

Annual GDP ARIMA(1,1,0) model Annual GDP External Demand Model

QRGDP Trend + AR(1) + Seasonal dummies QRGDP AR(1,4) + SARIMA(0,1,0,4)

Page 18: National Macroeconomic Forecasting And Policy Simulation ... · •Using the Quartely Real GDP series (2003Q1 to 2015Q2) •Model using information on its own past values •AR (1

NGDP Forecasting method

GDP (EXP)

C

Imports of Household Consumption Goods

G

Wage bill budget and forecast

0.75

Intermediate 

consumption costs 

budget and forecast

0.25 

I

Imports of capital goods

0.6

Imports of building material 

0.3

Govt Capital Expenditure budget and forecast

0.1

X (goods)

Exports of goods

1

X (services)

Tourism Bednights

1

M

Imports of goods

0.9

Wage bill budget and forecast

0.05

Govt Capital Expenditure budget and forecast

0.05

2014 GDP

2014 GDPExpenditure breakdowns

Forecast growth for indicators 2015‐2018

Indicator Weights 

Page 19: National Macroeconomic Forecasting And Policy Simulation ... · •Using the Quartely Real GDP series (2003Q1 to 2015Q2) •Model using information on its own past values •AR (1

Nominal GDP Series‐ October 2014 update 

‐10%

‐5%

0%

5%

10%

15%

20%

25%

30%

35%

0

10000

20000

30000

40000

50000

60000

70000

2001 2002 2003 2004 2005 2006 2007 2008 2009 2010 2011 2012 2013 2014 2015 2016 2017

Growth

MVR

 millions

Source : National Bureau of Statistics (October 2014)

Page 20: National Macroeconomic Forecasting And Policy Simulation ... · •Using the Quartely Real GDP series (2003Q1 to 2015Q2) •Model using information on its own past values •AR (1

Issues with GDP forecasting

• Disconnected methods to forecast RGD and NGDP• Not a critically scrutinized model

Page 21: National Macroeconomic Forecasting And Policy Simulation ... · •Using the Quartely Real GDP series (2003Q1 to 2015Q2) •Model using information on its own past values •AR (1

Forecasting ModelsTourist Arrivals

Page 22: National Macroeconomic Forecasting And Policy Simulation ... · •Using the Quartely Real GDP series (2003Q1 to 2015Q2) •Model using information on its own past values •AR (1

Medium term forecasts

• A global tourism elasticity‐based model for medium term forecasts• Maldives is hyper‐sensitive to global tourism, with an estimated elasticity of 184 percent

Page 23: National Macroeconomic Forecasting And Policy Simulation ... · •Using the Quartely Real GDP series (2003Q1 to 2015Q2) •Model using information on its own past values •AR (1

Medium term model

• LOG(ARR_MDV) = C(1) + C(2)*LOG(ARR_GLOBAL) + C(3)*SHOCK + AR(1)

• Global arrivals forecasts from UNWTO is included. 

Page 24: National Macroeconomic Forecasting And Policy Simulation ... · •Using the Quartely Real GDP series (2003Q1 to 2015Q2) •Model using information on its own past values •AR (1

Within year forecast

• Temporal (or seasonal) based model for within year forecasts• Seasonal patterns apparent and consistent• High dependency on European tourists, peak season in Maldives corresponds to European winter months of December‐February. 

• However, since of late China has emerged as the single largest source of tourists and also the one which has shown the strongest growth. Interestingly, Chinese arrivals also reflect strong seasonal features but with effects opposite to that of the European segment. 

Page 25: National Macroeconomic Forecasting And Policy Simulation ... · •Using the Quartely Real GDP series (2003Q1 to 2015Q2) •Model using information on its own past values •AR (1

Within year (short term) model

• Calculate weights for main markets, namely, China, Europe and RoW.

• Run AR models for each market• Multiply forecasts from AR models with weights to obtain short term monthly forecasts

Page 26: National Macroeconomic Forecasting And Policy Simulation ... · •Using the Quartely Real GDP series (2003Q1 to 2015Q2) •Model using information on its own past values •AR (1

Issues and updates

• Both models have relatively large standard errors.• Shock variable (a binary variable) is exogenous to the model and is captured for only 1 period. This is not very robust and a better approach would be to have some indication of duration of a shock through an appropriate formulation.

• Models performed poorly in recent years.• New research to update the model on‐going.

Page 27: National Macroeconomic Forecasting And Policy Simulation ... · •Using the Quartely Real GDP series (2003Q1 to 2015Q2) •Model using information on its own past values •AR (1

Forecasting ModelsRevenues

Page 28: National Macroeconomic Forecasting And Policy Simulation ... · •Using the Quartely Real GDP series (2003Q1 to 2015Q2) •Model using information on its own past values •AR (1

Revenue Forecasting

• Historically done by collecting agencies.• New Medium Term Economic Framework (MTEF) being built. Same models are incorporated to the MTEF.

Page 29: National Macroeconomic Forecasting And Policy Simulation ... · •Using the Quartely Real GDP series (2003Q1 to 2015Q2) •Model using information on its own past values •AR (1

• Over 90% of 2014 MIRA revenue comprise of 7 revenue codes:

• GST (Tourism and Non‐Tourism Sectors)

• BPT• Tourism Land Rent• Tourism Tax / Green Tax• Lease Period Extension Fee• Bank Profit Tax• Airport Service Charge

2.41

4.56

7.15

8.98

11.50

0

2

4

6

8

10

12

14

2010 2011 2012 2013 2014

MVR

 billions

Others Airport Service Charge

Bank Profit Tax Lease Period Extension Fee

Tourism Tax Tourism Land Rent

BPT GST

Total

MIRA Revenue

Page 30: National Macroeconomic Forecasting And Policy Simulation ... · •Using the Quartely Real GDP series (2003Q1 to 2015Q2) •Model using information on its own past values •AR (1

Tourism Sector GST• Bednights information and forecast – from the Ministry of Tourism

• Spending Rate• Calculate the ‘consumption’ of the tourism sector

• Equals to TGST collection / TGST rate• Calculate Spending per tourist per bednight (spending rate)

• Tourism sector consumption / bednights• Estimate the spending rate for the projection period

• Take a moving average of the past years’ spending rate• Increase it by a rate indicative to the growth in past year’s spending rate and inflation

• TGST revenue = Spending Rate x Bednights x  TGST Rate

150

200

250

300

350

400

450

500

Jan Feb Mar Apr May Jun Jul Aug Sep Oct Nov Dec

Spending Rate

2011 2012 2013 20142015 2016 2017

Page 31: National Macroeconomic Forecasting And Policy Simulation ... · •Using the Quartely Real GDP series (2003Q1 to 2015Q2) •Model using information on its own past values •AR (1

Non‐Tourism Sector GST• Calculate the consumption subject to GST in the non‐tourism sector

• Equals to GGST collection / GGST Rate

• Analyse the consumption trends

• Analyse GDP and inflation• GDP forecast not available at the time of projection• Estimate a crude growth rate for the real GDP based on past growth rates• Estimate a crude inflation figure for the projection period based on past figures• Estimate nominal GDP growth based on inflation and real GDP growth rate

• Estimate the consumption for the projection period based on moving average of past consumption levels, increased by nominal GDP growth rate.

• Additional weight may be necessary to reflect growing compliance 

• GGST revenue = Consumption x GGST Rate

Page 32: National Macroeconomic Forecasting And Policy Simulation ... · •Using the Quartely Real GDP series (2003Q1 to 2015Q2) •Model using information on its own past values •AR (1

Airport Service Charge

• Tourist arrivals estimates are obtained from the Ministry of Tourism• Calculate the share of foreign departures based on information obtained from Airport operators

• Share of foreigners in the total departures: 88.4%• Share of tourists among foreign departures: 92.1%

• Estimate foreign departures• Tourist arrivals divided by ‘share of tourists among foreign departures’

• Estimate total departures• Foreign departures divided by ‘share of foriegners in the total departures’

• Estimate Maldivian departures to abroad• Total departures minus foreign departures

• Multiply with respective ASC rates • USD 25 for foreigners and USD 12 for Maldivians

• Airport Service Charge = ASC from Foreigners + ASC from Maldivians

Page 33: National Macroeconomic Forecasting And Policy Simulation ... · •Using the Quartely Real GDP series (2003Q1 to 2015Q2) •Model using information on its own past values •AR (1

Economic Forecasting in Maldives

• Various models, various institutions and various software.• The teams involved in forecasting have close dialogue and work together on some models (eg: new tourist arrivals model)