FOCUSED AND QUICK (FAQ) Issue 40
How consumption responds to the interplay between inflation and income?
Pranee Sutthasri
Pensiri Kangvonkit
Summary
Consumption is mainly determined by real income growth and inflation, but for the same percentage increase in the two factors, the positive impact of income is found to be stronger than the negative impact from inflation. Meanwhile, the effect of real interest rate on consumption is small.
For 2011, despite the stronger acceleration in inflation relative to real income growth, the positive income effect on total consumption still outweighs the negative price effect. This means that the growth momentum for overall consumption is likely to remain respectable and that consumption will maintain its role as a key driver of the Thai economy for the rest of this year. Interestingly, this comes primarily from the expected acceleration in durable goods consumption while the growth of nondurable goods consumption in 2011 may slow down from 2010.
It is common knowledge that the key determinants of consumption are inflation and real income growth. Therefore, fear of a drop in consumption often arises when inflation
Table 1: Mean and standard deviation of key macro variables' year-on-year growth rates
Year PCE GDP HeadlineInflation
Avg. s.d. Avg. s.d. Avg. s.d. 1994-2010 3.3 5.0 3.9 5.3 3.4 2.6 1994-2000 2.6 7.3 3.1 7.0 4.6 2.8 2001-2010 3.8 2.3 4.4 3.7 2.6 2.2 Source: NESDB, authors' calculation
With the recent pick-up in inflation, the same concern has once again reemerged. This article thus seeks to clarify whether or not that fear is justified, using an econometric tool which leads in turn to the analysis of the consumption growth momentum going forward.
Policy makers in most countries pay close attention to the development of PCE because it is regarded as a key contributor to economic growth. Just as in many other economies, PCE constitutes the largest component of domestic demand for Thailand .
During 1994-2010, PCE was growing at an average rate of 3.3% per year. With a relatively stable share of GDP at around 50-55% coupled with its resilient nature as reflected by a lower standard deviation (s.d.) value than GDP, PCE plays a crucial role both in determining Thailand's economic growth and in stabilizing the growth path.
PCE components can be grouped into non-durable and durable goods consumption. Durable goods consumption*(1)(with a share of around 30% of PCE) is more volatile and, on average, demonstrates higher growth vis-เ-vis non-durable goods consumption*(2) and real GDP. This is consistent with the economic intuition that durable goods represent discretionary consumption whereas non-durable goods represent necessity consumption.
Now let's take a closer look at the interplay between the main determinants of PCE, namely inflation and real income growth. To obtain a clearer picture, we use the Error Correction Model (ECM) to estimate 3 equations for PCE, durable goods consumption, and non-durable goods consumption, respectively. The ECM is chosen because it provides both the long-run (LR) equilibrium relationship and the short-run (SR) dynamic adjustment towards equilibrium via the error correction term.
The data used are NESDB's quarterly PCE figures (denoted as CPR) from 1993-2010, along with the two components of PCE: durable goods consumption (CPR1) and nondurable goods consumption (CPR2). In compliance with the consumption theories, consumption is specified as a function of real disposable income (YD)3, headline inflation (HCPI), real minimum lending rate (RMLR)4, and real financial wealth, which is proxied by the SET index (SETR)5. The data are seasonallyadjusted and expressed in natural logarithms (ln) where appropriate.
The results comply with our expectation (see Appendix 6). Real income growth has a positive and highly significant role in determining all types of consumption. Meanwhile, inflation has a negative and significant impact on all types of consumption, but the impact on durable goods consumption is found to be statistically less significant though comparable in size to that of nondurable goods consumption.
Interest rate is found to have a negative effect on all types of consumption. The effect, however, is small in the long run while statistically insignificant in short-run. This might be explained by the offsetting influences of income (wealth) and substitution effects through the interest rate variable.Wealth , on the other hand, seems to affect only durable consumption with a positive sign, and as found in many other studies for the case of Thailand, the impact is rather small and not very significant at the aggregate level.
The variables used to capture habit persistence in consumption (CPR_SA(-1)), (CPR2_SA(-1)) also appear to have a positive role in determining current PCE and nondurable goods consumption, but not in the case of durable consumption.
The error correction terms have the expected negative coefficients which are significant at the 1% confidence level for all equations. Durable goods consumption appears to adjust more quickly than nondurable goods consumption and PCE, as reflected by the speed of adjustment of about 44% per period, which means 44% of the disequilibria in durable goods consumption is corrected per quarter.
As seen from the estimated results, the main drivers of consumption are real income growth and inflation while the effects from other variables are either insignificant or small. According to Table 2, for the same percentage increase in real income growth and inflation, the positive income effect more than offsets the negative impact from inflation when occurring simultaneously.
elasticity elasticity
PCE 0.84 -0.59 Durable 1.68 -0.77 Non-durable 0.53 -0.52 Source: Authors’ calculation aElasticity = SR coefficient + ((LR coefficient) *(coefficient of the ECM lagged term))However, the degree of sensitivity varies from one type of consumption to another. Non-durable goods consumption appears to be more resilient to either real income or price shock, as indicated by a smaller absolute value of income and price elasticity compared to durable goods consumption. This is consistent with the general observation that non-durable goods consumption is more stable than durable goods consumption.
Looking at the relative size of income and price elasticity alone is not enough to conclude whether the currently rising inflation will significantly dent the momentum of consumption going forward. Here we take the rate of acceleration in inflation, as well as the likely trajectory of other variables, into consideration.
By the end of this year, real MLR is expected to remain fairly low at around 2.3 to 2.8%.*(8) That coupled with the modest net impact of interest rate on consumption implies that the current upward interest rate cycle is not going to stall consumption.
Meanwhile, we translate the BOT's forecast of year-on-year real income growth and inflation in 2011 into quarter-on-quarter annualized rates.*(9) As shown in Table 3, both real income and inflation are on the accelerating path from last year, with inflation accelerating faster than real income as reflected by the difference in the annualized QoQ rates of 1.6% and 2.1% for real income and inflation, respectively.
YoY QoQ QoQa YoY QoQ QoQa
Real GDP 7.8 1.0 4.1 4.1 1.4 5.7 growth Headline 3.3 0.7 2.8 3.9 1.2 4.9 inflation Source: Year-on-year forecast for 2011 from the BOT, Inflation Report (April 2011); quarter-on-quarter forecast from the authors’ own calculation based on the corresponding year-on-year number; “a” denotes an annualized rateDespite a stronger acceleration of inflation relative to real income growth, the larger income elasticity relative to the price elasticity still makes the positive income effect outweigh the negative price effect for PCE (total consumption). This means that the QoQ annualized growth rate of PCE in 2011 is likely to be stronger than in 2010 , despite the pick-up in inflation, and that PCE will maintain its role as a key driver for the Thai economy. More interestingly, such positive net effect comes primarily from the expected acceleration in durable goods consumption while the growth of nondurable goods consumption in 2011 may slow down from 2010.
The authors are grateful to Roong Poshyananda Mallikamas, Pornvipa Tangcharoenmonkong, and Jeerapan O-lanthanasate for their very useful guidance and support.
*(1) Including metal products, machinery and equipment, and other transportable goods.
*(2) Including food, beverage, tobacco, textile, agricultural products, and services.
*(3) Real disposable income = real GDP*(1-effective tax rate), where effective tax rate = personal income tax divided by nominal GDP
*(4) RMLR = (MLR — headline inflation)/(1 + headline inflation)
*(5) SETR = SET index divided by headline CPI
*(6) Consumer Confidence Index (CCI) and consumer credits were included in some specifications to capture the plausible impact of consumer confidence and credit access, respectively, but their effects were found to be insignificant at the aggregate level.
*(7) The substitution effect occurs as higher interest rate increases the cost of borrowing and encourages saving, which leads in turn to a reduction in current consumption, whereas the income effect occurs when higher interest rate increases household’s financial wealth (only if they have positive net position), which has a positive effect on consumption.
Senior Economist
Senior Economist
Ms. Pensiri
Kangvonkit conomist
Domestic Economy Department
Craigwell, Roland C., and Llewyn L. Rock. 1995. "An aggregate consumption function for Canada: a cointegration approach." Applied Economics , 27(2): 239-249.
Casuax, Suzanne, and Pierre Ecochard. 2011. "Private consumption in France-Stubbornly high or responsive to determinants?." ECFIN Country Focus , 8(1).
Friedman, Milton. 1957. "A Theory of the Consumption Function." Princeton University Press.
Johansen, Sren, and Katarina Juselius. 1990. "Maximum Likelihood estimation and inference on cointegration-with applications to the demand for money," Oxford Bulletin of Economics and Statistics , 52: 169-210. 1936. "The General Theory of Employment, Interest and Money."
Keynes, John Meynard. Macmillan, London.
Modigliani, Franco, and Richard Brumberg. 1954. "Utility analysis and the consumption function: An interpretation of cross-section data, in Post-Keynesian Economics." ed. K. K. Kurihara. New Brunswick, NJ: Rutgers University Press.
Rungcharoenkitkul, Phurichai. 2011. "Wealth Effects and Consumption in Thailand." Bank of Thailand Discussion Paper 01/2011.
Saad, Wadad. 2011. "An Econometric Study of the Private Consumption Function in Lebanon." International Research Journal of Finance and Economics , 61: 29-41.
The World Bank. "Thailand's economic monitor." 1999-2011.
Source: Bank of Thailand