Firm Inflation Expectations and Monetary Policy in Uruguay - bvrie

Firm Inflation Expectations and Monetary Policy in Uruguay Gerardo Licandro Miguel Mello

006 - 2014

1688-7565

Firm Inflation Expectations and Monetary Policy in Uruguay Gerardo Licandro 1 ª1, Miguel Mello 2 b**, a Banco Central del Uruguay (Inveco), 777 Diagonal J.P. Fabini 11100 Montevideo, Uruguay b Banco Central del Uruguay (Inveco), 777 Diagonal J.P. Fabini 11100 Montevideo, Uruguay

Documento de trabajo del Banco Central del Uruguay 2014/006 Autorizado por: Gerardo Licandro

Abstract Using a novel monthly survey of firm inflation expectations for Uruguay from October 2009 to June 2013, this paper studies the impact of monetary policy on inflation expectations at the micro level. Using several panel data techniques we consistently find a negative and statistically significant relationship between monetary policy and inflation expectations. We also find a high level of inertia in expectations. Past inflation changes have a positive impact on inflation expectations, while exchange rate changes have a significant but low importance on expectations. We observe a negative link between inflation expectations and expected economic activity, potentially due to past experiences of a monetary financing of crisis. Contrary to intuition, there is no clear link between firm inflation expectations and the median assessment of experts published by the Central Bank.

JEL: C18, C82, E23, E24, J21, L16, L70 Keywords: Monetary transmission, inflation expectations, expectations channel.

* Correo electrónico: [email protected] ** Correo electrónico: [email protected]

1 2

We thank participants at the seminars of the Economics Department of the FCS- UDELAR, and the ORT University seminars, and Banco Central del Uruguay for valuable comments. We want to thank José Licandro, Diego Gianelli and an anonymous referee of the Research Network of CEMLA for their comments. Any remaining mistakes are our sole responsibility,

1

2

3

4

5

6

7

1

The IMF has collected data on inflation expectations of some private sector consultants over the years Uruguay has had programs the IMF. The dataset is unbalanced, has only a few opinions (at times only one), and has important data gaps. At the beggining information was collected yearly. To work with this data, Gelós and Rossi (2007) use interpolation methods to fill the gaps.

8

9

10

11

12

   

2

There are evident differences in the way volatility in the money market is carried to aggregates and interest rate, the interest rate stabilizes the demand shocks and the aggregates provide volatility to the interest rate, consumption and investment. However, impacts on relative volatility of interest rates and monetary aggregates aside, the sign of the effects of monetary policy shocks remain the same.

13

 



𝐼𝑀𝑡 = 𝑇𝑃𝑀𝑡 − 𝑇𝑁𝐼𝑡 𝑇𝑁𝐼𝑡 = 𝑟 + 𝜋 𝑒 𝑡 _18

𝐼𝑀𝑡

14

𝜋 𝑒 𝑡 _18

3

Since we used the median of the National Statistics Institute survey, it isn’t feasible to find herd effect as Borraz and Gianelli (2010) did since this survey isn’t public. 4 In a classical conception, where prices are flexible, the shocks that cause the cycle are real shocks, that affect capital accumulation during the steady state and the real interest rate. Within this environment and given price flexibility, monetary policy has no effects. In new keynesian models, which presents price rigidities, it is implied that offer conditions are relatively stable and that fluctuations arise from the demand side.{

15

𝐼𝑀𝑡2 = 𝑇𝑃𝑀𝑡 − [𝑇𝑁𝐼𝑡 + 0.5(𝑦𝑡 − 𝑦𝑡𝑝𝑜𝑡𝑒𝑛𝑡𝑖𝑎𝑙 ) + 1.5(𝜋𝑡 − 𝜋

5

𝑡𝑎𝑟𝑔𝑒𝑡

)]

In the appendix we show the graph for the three monetary stances used in this paper.

16

3

2

1

0

-1

-2

-3 2005

2006

2007

2008

2009

INST_MON_TC

2010 M_Y_TC

17

2011

2012

2013

4 3 2 1 0 -1 -2 -3 I

II

III

2009

IV

I

II

III

2010

IV

I

II

III

IV

2011 INST_MON_TC INF_ESP_12M_EMP_TC INF_ESP_12M_BCU_TC

18

I

II

III

2012

IV

I

II

2013

3

2

1

0

-1

-2

-3 2005

2006

2007

2008 EXP_ECO_TC

inf_esp_12m mediana_inf_esp_18m inf_esp_18m

2009

2010

2011

2012

INST_MON_TC

inf_esp_12m mediana_inf_esp_18m inf_esp_18m 1 0.1236 1 0.8926 0.1029 1

19

2013

𝑒 𝑒 𝜋𝑖,𝑡 = 𝜙𝑖 + ∑𝑘0 𝛼 . 𝜋𝑖,𝑡−𝑠 + ∑𝑘0 𝜌 . (𝐼𝑀𝑡−𝑠 ) + ∑𝑘0 𝜆 . 𝜋𝑡−𝑠 + ∑𝑘0 𝛿 . 𝑑𝑡−𝑠 + 𝛽. 𝑀𝑡 + 𝜀𝑖,𝑡

𝑒 𝜋𝑖,𝑡

𝐼𝑀𝑡−𝑠 𝜋𝑡−𝑠 𝑑𝑡−𝑠 𝜙𝑖

6

The business, macroeconomic and international variables used can be found in the Appendix.

20

21

8

We do not present in our tables the Moulton results because it is exactly the same as OLS, what it means that there is no aggregation bias in our models.

6

The models estimated can be found in the Appendix.

22

23

6

The models estimated can be found in the Appendix.

24

11

Licandro and Vicente (2006) explore the relationship between monetary and fiscal policy in the Uruguayan case.

25

Inflation Expectations Inertia whole period Inflation Expectations Inertia June 2013 Simple stance impact on expectations Taylor rule stance impact on expectations Stance with HP filter natura rate impact on expectations

26

OLS

Fix Effects

6.58 7.15

6.48 7.02

-0.28 -0.32 0

-0.32 -0.25 0

Arellano-Bond Arellano-Bond One Stage Two Stages 5.19 5.23 5.73 5.77 -0.34 -0.43 -0.51

-0.33 -0.42 -0.51

12

This means that in order to align firm inflation expectations with the center of the target range of the

Banco Central del Uruguay it would be necessary to make the monetary stance more contractive in about 7%. 13

This long run multiplier implies a parameter of 3 for the response of monetary stance to the increase in inflation expectations in a Taylor Rule.

27

28

Working

29

http://www.cinve.org.uy/wpcontent/uploads/2013/01/Negociacion-Colectiva.pdf

na decada de metas de inflación en la región.”

‐

30

31

Different monetary stance approximations used in our estimations 3

2

1

0

-1

-2 IV

I

2009

II

III

IV

I

2010

II

III

IV

2011 MSTR

I

II

III

IV

I

2012 MSHP

II 2013

SMS

Variables used in the panel data estimates infesp_12m

expected inflation for a 12-month horizon

infesp_1

first delay of the expected inflation within 12 months

infesp_2

second delay of the expected inflation within 12 months

infesp_3

third delay of the expected inflation within 12 months

infesp_18m

expected inflation for an 18-monthhorizon

var_ctos_12m

expected cost variation expected inflation for a 12-month horizon

var_ctos_18m

expected inflation for an 18-month horizon

inst_mon

Monetary policy stance created through equations (1) and (2)

inst_1

first lag of the monetary policy stance

inst_2

second lag of the monetary policy stance

inst_3

third lag of the monetary policy stance

inst_tr

Monetary policy stance created with a Taylor rule, equation (3)

inst_tr_1

first lag of the monetary policy stance created with the Taylor rule

inst_tr_2

second lag of the monetary policy stance created with the Taylor rule

inst_tr_3

third lag of the monetary policy stance created with the Taylor rule

inst_tn_hp

Monetary policy stance created with a HP Filter for the natural interest rate

inst_tn_hp_1

first lag of the monetary policy stance created with a HP Filter for the natural

32

interest rate inst_tn_hp_2

call

second lag of the monetary policy stance created with a HP Filter for the natural interest rate third lag of the monetary policy stance created with a HP Filter for the natural interest rate interbank call rate

pib

gross domestic product

tc

exchange rate

deva

Uruguayan pesos annual depreciation rate relative to the dollar

tcr

real exchange rate

icc

confidence index of consumers

exp_cons_pai s exp_inf_bcu

consumer expectations relative to economic evolution of the country

exp_pib_bcu

ubi

median of the expectations of the GDP evolution drawn up by the Central Bank of Uruguay on its monthly economic expectation survey economic expectation of business owners as from the survey of the Chamber of Industries of Uruguay firm expectation of business owners as from the survey of the Chamber of Industries of Uruguay median of tax expectations drawn up by the Central Bank of Uruguay on its monthly economic expectation survey Uruguay country risk

inf

anual inflation rate

w_real

real wage

ims

median wage index

desempleo

annual unemployment rate

rti

relationship of terms of trade

embi

embi+ Uruguay

gap_y

Gap of the product as a diversion from a Hodrick and Prescott tendency

ctopib

GDP growth rate

tar_pub

Variation rate of public rates

inst_tn_hp_3

exp_eco exp_emp exp_fiscal

median of the Central Bank of Uruguay expected inflation survey

Ordinary Least Squares Linear regression

Number of obs F( 8, 12914) Prob > F R-squared Root MSE

infesp_12m

Coef.

infesp_1 infesp_2 infesp_3 inst_2 inf_1 exp_eco exp_inf_bcu deva _cons

.5605554 .2046169 .1293337 -.1736072 .1528616 -.4020178 -.0129869 .0047522 -.7211809

Robust Std. Err. .0179994 .0212715 .0190502 .0277431 .0134337 .1885009 .0408234 .0011232 .2606691

t

33

31.14 9.62 6.79 -6.26 11.38 -2.13 -0.32 4.23 -2.77

P>|t| 0.000 0.000 0.000 0.000 0.000 0.033 0.750 0.000 0.006

= 12923 = 2444.17 = 0.0000 = 0.7590 = .80526

[95% Conf. Interval] .5252739 .1629217 .0919926 -.2279878 .1265296 -.7715075 -.0930069 .0025506 -1.232131

.5958369 .2463121 .1666748 -.1192267 .1791937 -.0325281 .067033 .0069538 -.2102309

Fixed Effects Model Fixed-effects (within) regression Group variable: ine

Number of obs Number of groups

= =

12923 529

R-sq:

Obs per group: min = avg = max =

1 24.4 42

within = 0.4639 between = 0.9515 overall = 0.7458

corr(u_i, Xb)

F(8,528) Prob > F

= 0.7120

= =

335.66 0.0000

(Std. Err. adjusted for 529 clusters in ine) Robust Std. Err.

infesp_12m

Coef.

t


.4501237 .1378325 .0471683 -.1278855 .1562438 -.7844812 .1082978 .0042073 .691257

.0199227 .0210155 .0143448 .0235185 .015707 .1577779 .0390259 .0011054 .2525182

sigma_u sigma_e rho

.54012893 .77456954 .32717301

(fraction of variance due to u_i)

22.59 6.56 3.29 -5.44 9.95 -4.97 2.78 3.81 2.74

P>|t| 0.000 0.000 0.001 0.000 0.000 0.000 0.006 0.000 0.006

[95% Conf. Interval] .4109862 .0965482 .0189885 -.1740868 .125388 -1.094431 .0316327 .0020358 .1951933

.4892612 .1791169 .0753481 -.0816842 .1870997 -.4745318 .1849629 .0063788 1.187321

.

Random Effects Model Random-effects GLS regression Group variable: ine

Number of obs Number of groups

= =

12923 529

R-sq:

Obs per group: min = avg = max =

1 24.4 42

within = 0.4542 between = 0.9587 overall = 0.7588

corr(u_i, X)

Wald chi2(8) Prob > chi2

= 0 (assumed)

= =

9983.75 0.0000

(Std. Err. adjusted for 529 clusters in ine) Robust Std. Err.

infesp_12m

Coef.

z


.540374 .1944679 .1174183 -.1674245 .1542208 -.4686147 .0063285 .0047546 -.4959281

.0179575 .0217314 .0143441 .0230142 .0151758 .157488 .0384546 .0010233 .2392888

sigma_u sigma_e rho

.12233232 .77456954 .02433669

(fraction of variance due to u_i)

30.09 8.95 8.19 -7.27 10.16 -2.98 0.16 4.65 -2.07

34

P>|z| 0.000 0.000 0.000 0.000 0.000 0.003 0.869 0.000 0.038

[95% Conf. Interval] .505178 .1518751 .0893044 -.2125314 .1244768 -.7772855 -.0690411 .0027489 -.9649255

.5755701 .2370607 .1455322 -.1223176 .1839648 -.1599439 .0816981 .0067603 -.0269307

Arellano/ Bond model in one stage Dynamic panel-data estimation, one-step system GMM Group variable: ine Time variable : fecha Number of instruments = 500 Wald chi2(8) = 1153.03 Prob > chi2 = 0.000 infesp_12m

Coef.


.4754623 .1395857 .1874863 -.146728 .1991429 -.863634 .166095 .0061208 -1.446348

Number of obs Number of groups Obs per group: min avg max Robust Std. Err.

z

.0375451 .0410212 .0274096 .0349289 .0209245 .2481966 .0572149 .0018863 .4046236

12.66 3.40 6.84 -4.20 9.52 -3.48 2.90 3.24 -3.57

P>|z| 0.000 0.001 0.000 0.000 0.000 0.001 0.004 0.001 0.000

= = = = =

12923 529 1 24.43 42

[95% Conf. Interval] .4018753 .0591857 .1337644 -.2151875 .1581316 -1.35009 .0539558 .0024238 -2.239396

.5490493 .2199857 .2412082 -.0782685 .2401542 -.3771776 .2782342 .0098178 -.6533004

Instruments for first differences equation Standard D.var_ctos_18m GMM-type (missing=0, separate instruments for each period unless collapsed) L(1/12).var_ctos_12m Instruments for levels equation Standard var_ctos_18m _cons GMM-type (missing=0, separate instruments for each period unless collapsed) D.var_ctos_12m Arellano-Bond test for AR(1) in first differences: z = Arellano-Bond test for AR(2) in first differences: z = Sargan test of (Not robust, Hansen test of (Robust, but

overid. restrictions: chi2(491) =2678.30 but not weakened by many instruments.) overid. restrictions: chi2(491) = 493.11 weakened by many instruments.)

-7.38 2.10

Pr > z = Pr > z =

0.000 0.036

Prob > chi2 =

0.000

Prob > chi2 =

0.465

Difference-in-Hansen tests of exogeneity of instrument subsets: GMM instruments for levels Hansen test excluding group: chi2(449) = 468.25 Prob > Difference (null H = exogenous): chi2(42) = 24.86 Prob > iv(var_ctos_18m) Hansen test excluding group: chi2(490) = 493.05 Prob > Difference (null H = exogenous): chi2(1) = 0.06 Prob >

35

chi2 = chi2 =

0.256 0.984

chi2 = chi2 =

0.453 0.807

Arellano/ Bond model in two stages Dynamic panel-data estimation, two-step system GMM Group variable: ine Time variable : fecha Number of instruments = 500 Wald chi2(8) = 1141.63 Prob > chi2 = 0.000 infesp_12m

Coef.


.4754706 .1400988 .1876232 -.1420387 .1977725 -.8506022 .1638629 .0059895 -1.410897

Number of obs Number of groups Obs per group: min avg max Corrected Std. Err. .0375434 .040946 .0275919 .034493 .020552 .2500994 .056999 .0018765 .4095289

z 12.66 3.42 6.80 -4.12 9.62 -3.40 2.87 3.19 -3.45

P>|z| 0.000 0.001 0.000 0.000 0.000 0.001 0.004 0.001 0.001

= = = = =

12923 529 1 24.43 42

[95% Conf. Interval] .4018869 .0598462 .133544 -.2096438 .1574914 -1.340788 .0521468 .0023116 -2.213559

.5490542 .2203514 .2417023 -.0744337 .2380537 -.3604165 .2755789 .0096674 -.6082356

Instruments for first differences equation Standard D.var_ctos_18m GMM-type (missing=0, separate instruments for each period unless collapsed) L(1/12).var_ctos_12m Instruments for levels equation Standard var_ctos_18m _cons GMM-type (missing=0, separate instruments for each period unless collapsed) D.var_ctos_12m Arellano-Bond test for AR(1) in first differences: z = Arellano-Bond test for AR(2) in first differences: z = Sargan test of (Not robust, Hansen test of (Robust, but

overid. restrictions: chi2(491) =2678.30 but not weakened by many instruments.) overid. restrictions: chi2(491) = 493.11 weakened by many instruments.)

-6.74 1.38

Pr > z = Pr > z =

0.000 0.167

Prob > chi2 =

0.000

Prob > chi2 =

0.465

Difference-in-Hansen tests of exogeneity of instrument subsets: GMM instruments for levels Hansen test excluding group: chi2(449) = 468.25 Prob > Difference (null H = exogenous): chi2(42) = 24.86 Prob > iv(var_ctos_18m) Hansen test excluding group: chi2(490) = 493.05 Prob > Difference (null H = exogenous): chi2(1) = 0.06 Prob >

36

chi2 = chi2 =

0.256 0.984

chi2 = chi2 =

0.453 0.807

Ordinary Least Square with Monetary Stance HP

Fixed Effects model with Monetary Stance HP

37

Arellano/ Bond model in one stage with monetary stance HP

38

Arellano/Bond model in two stages with monetary stance HP

39

Ordinary Least Squares with monetary stance with Taylor rule

Fixed Effects with monetary stance with Taylor rule

40

Arellano/Bond in one stage with monetary stance with Taylor rule

41

Arellano/Bond in two stages with monetary stance with Taylor rule

42