Monetary Policy Transmission in an Emerging Market: �Financial Friction Channel VS Interest-Rate Channel ��Lorenzo Menna, Martin Tobal,¹ and Alejandro Werner^2�

1

1/ Director of Macro-Financial Risk Analysis at the Central Bank of Mexico (Banco de Mexico). 2/ Georgetown Institute.

The views here are my own and do not necessarily represent those of the Bank of Mexico or its Board of Governors.

Uso Público� �Información de acceso público.

1

1. Introduction

• When talking about monetary policy transmission, macroeconomist go back to the concept of financial frictions.
• Financial frictions: obstacles that prevent credit from going from potential suppliers to potential demanders.
• Financial frictions:
• are linked to limited available information for assessing borrowers’ risk of default.
• This limited information reflects, in part, that firms themselves have business records and assets that are hard to verify.
• Thus, banks impose collateral or networth requirements.
• That is, financial frictions reduce the value of credit transactions, which can be seen in 2 mainstream models:
1. Costly-state-verification framework (Bernanke & Gertler, 1989). Banks face monitoring costs to analyze business performance, and the need to pay for these costs diminishes the value of credit transactions. Firms respond by financing their investments in part with internal funds;
2. Collateral-constraint framework (Kiyotaki and Moore, 1997). Banks impose collateral to protect themselves against the risk of default, but little availability and enforceability of collateral limits this mechanism.
1. Thus, frictions are higher where: (i) there is less information; & (ii) collateral is more limited or harder to enforce.

​

Monetary Policy Transmission in an Emerging Market: The Financial- Friction Channel VS The Interest-Rate Channel

Uso Público� �Información de acceso público.

2

1. Introduction

• Given this, it is easy to see financial frictions are higher in Emerging Market and Developing Economies (EMDEs) because:
1. In these countries, banks have less information because:
• Credit bureaus are less developed (Djankov et al., 2007);
• Disclosure standards are weaker, which makes it harder to compare projects across firms (La Porta et al., 1998).
• Firms frequently lack traceable transactions (La Porta and Shleifer, 2014).
2. In these countries, the availability and enforceability of collateral is more limited.
• Bankruptcy laws and courts make it harder for banks to enforce collateral rights (Levine, 1998, 1999).
1. These features of EMDEs increase the obstacles between banks and borrowers and thus financial frictions.
2. Given this and that financial frictions are critical in monetary policy transmission, a natural question arises: is the transmission of monetary policy different in EMDEs?
3. However, despite the importance of this question, the scarcity of granular data in EMDEs has prevented us from fully understanding how financial frictions shape monetary policy transmission in these economies.

​

Monetary Policy Transmission in an Emerging Market: The Financial- Friction Channel VS The Interest-Rate Channel

Uso Público� �Información de acceso público.

3

1. Introduction

• We fill this gap: we use a highly granular dataset to study monetary policy transmission in an EMDE: Mexico.
• The granularity of these data allows us to consider 2 transmission mechanisms:
1. Financial Frictions Channel (FFC), whose strength depends on the CHARACTERISTICS of BORROWING FIRMS.
• When monetary policy tightens, the value of networth and collateral decrease, thus making extending credit riskier for banks.
• Credit may decline first for firms about which less information is available and that, as a result, require greater net worth and collateral; for example, young firms with short credit histories from which lenders can extract only limited signals.
2. Interest Rate Channel (IRC), whose strength depends on SECTOR DURABILITY.
• Durable goods provide services over time; thus, they are closer to financial assets because both help move resources into the future.
• When monetary policy tightens, the return on financial assets increases, making these more attractive relative to durable goods.
• Thus, credit may fall first for firms producing durable goods.
• Hence, to test for the FFC and the IRC, one needs information on firms,’ characteristics, notably, their age, and the sector where they operate.

Monetary Policy Transmission in an Emerging Market: The Financial- Friction Channel VS The Interest-Rate Channel

Uso Público� �Información de acceso público.

4

1. Introduction

• We use an administrative dataset collected by the central bank and financial regulator of Mexico.
• It contains loan-level information on the universe of loans granted by private commercial banks to private firms.
• Aggregated at the firm-month level, it yields over 10 million observations for more than 225,000 firms, providing us with enough statistical power to study monetary policy transmission in an EMDE.
• It provides large heterogeneity across characteristics of borrowing firms and the sectors where they operate that we use it to test the financial-frictions and the interest-rate channels: we will not only assess these channel but also compare their relative strength.
• However, it does not have information on employees; thus, we use a public, external database on social security records at the regional level to gain preliminary insights into the effect of monetary policy on employment.
• We also use an additional public dataset on banks’ balance sheets to investigate the role of bank capitalization and provide further support for the financial frictions channel.
• In identifying monetary policy shocks, we use a high-frequency identification method: we calculate changes in expected interest rates within short windows around the moments Banxico announces its monetary policy decisions.
• We introduce this shock in a panel local projection model (Jordà, 2005) to estimate responses to monetary policy.

Monetary Policy Transmission in an Emerging Market: The Financial- Friction Channel VS The Interest-Rate Channel

Uso Público� �Información de acceso público.

5

Uso Público� �Información de acceso público.

6

2.1 Firm-level Data

• We use an loan-level administrative dataset covering the universe of loans granted by private banks to private firms.
• We aggregate loans at the firm-level and obtain a panel of over 10,548,362 firm-month observations.
• To examine the Financial Friction Channel, we use 3 firms-characteristics that capture how much information is available for credit-risk assessment and how strong collateral requirements are:
• FIRM-AGE: Young firms have shorter credit histories from which to extract signals.
• We follow Cloyne et al. (2023) and classify a firms as: young (0–10 years); mature (10–20 years); or old (older).
• FIRM-SIZE: Small firms have business records harder to verify.
• We follow the central bank’s flagship reports and classify a firm as: small (if it has not received a loan exceeding 100 million Mexican pesos (in 2018 constant prices)); or large (if it has received a loan over that threshold).
• FIRM-RECENT DELINQUENCY: Firms with recent defaults are perceived as riskier.
• We classify a firm as: a recent defaulter (if it defaulted the previous year); non-recent defaulter (otherwise).

​

Monetary Policy Transmission in an Emerging Market: The Financial- Friction Channel VS The Interest-Rate Channel

Uso Público� �Información de acceso público.

7

2.1 Firm-level Data

• Our prefered measure is firm-age because:
• It is the “most exogeneous” to both monetary policy and the business cycle.
• Existing studies in the literature (for AEs) also use age (Cloyne et al., 2023; Durante et al., 2022, & so on).
• Age may be even more relevant in EMDEs, where traditional mechanisms to learn about firms’ past behavior are less effective.
• To examine the Interest Rate Channel, we classify firms according to the durability of the sector in which they operate and, based on the North American Industry Classification (NAICS), group them into the same 4 groups as Durante et al. (2022):
• “Construction:” all 3-digit subsectors under the 2-digit NAICS code 23 (labeled as Construction);
• “Durable Manufacturing:” part of the 3-digit subsectors under NAICS codes 31–33 (Manufacturing Industries)
• “Non-Durable Manufacturing:” the other part of these 3-digit subsectors.
• “Services:” all remaining 3-digit subsectors are indeed services.

Monetary Policy Transmission in an Emerging Market: The Financial- Friction Channel VS The Interest-Rate Channel

Uso Público� �Información de acceso público.

8

2.3 Bank-Firm Level and Employment Data

• In the bank-level analysis, we use public data from the financial regulator to assess the role of bank capitalization.
• Lower capitalization reduces a bank’s ability to absorb losses, limiting its risk-taking capacity.
• For these banks, the financial frictions channel should operate more strongly.
• Thus, to provide further evidence for this channel, we:
• Link firms in firm-level credit dataset to their lending banks;
• Retrieve information from the financial regulator’s website 2 measures of banks’ capitalization (the capital adequacy (ICAP) and core capital ratios) and assign them to each of the links.
• We run the same empirical exercise within 2 subsamples of the “more-” and the “less-” capitalized banks.
• To gain insights into the employment effects, we use a public dataset from the Social Security Institute (IMSS).
• We group the employment data into municipalities and run a municipality level analysis.
• For this purpose, we classify municipalities depending on: (b) the share of young, mature and old firms in their credit; (b) the share of firms in sectors with different durability in their total credit.
• The employment response to the monetary policy shock based on the 1^st classification links to the FFC, while the the response based on the 2^nd classification links to the IRC.

Monetary Policy Transmission in an Emerging Market: The Financial- Friction Channel VS The Interest-Rate Channel

Uso Público� �Información de acceso público.

9

Uso Público� �Información de acceso público.

10

3.2 Empirical Framework

Monetary Policy Transmission in an Emerging Market: The Financial- Friction Channel VS The Interest-Rate Channel

Uso Público� �Información de acceso público.

11

Uso Público� �Información de acceso público.

12

4 Results: Average Effect

• The results in terms of sign, timing, significance and size are in line with those found in the literature.
• The sign of the impact is as expected, negative, i.e., a contraction in monetary policy reduces credit growth.
• The effect does not become statistically significant until a year after the shock, consistent with the fact that obtaining credit takes time.
• It persists for the whole second year.
• At month t+23, it diminishes 0.2 percentage points, so a 25 basis points surprise leads to a 5% contraction in credit growth.

Average Firm Level Credit Growth Response to Monetary Policy Shocks

Source: Authors’ calculations based on data from Banco de México.

Notes: Effect of a 1-basis-point monetary shock on credit growth. Shaded areas represent 90 and 95 percent confidence intervals. Errors clustered at firm and month level.

Monetary Policy Transmission in an Emerging Market: The Financial- Friction Channel VS The Interest-Rate Channel

Uso Público� �Información de acceso público.

13

4.1 The Financial-Frictions Channel: Firm Age

• The effect varies across firms with different characteristics, specifically firm-age: monetary policy reduces credit growth earlier for young firms than for mature (≈15 months) and than for old firms (≈16 months), with the impacts persisting for the whole second year.
• The magnitude is also larger (roughly twice as large) for young firms. At month t+16 after the shock, when the impact is significant for all groups of firms: credit growth diminishes by about 0.15 pp for young firms, compared to 0.07 pp for mature and 0.06 pp for old firms.
• An earlier and stronger response for young firms provides support for the Financial Friction Channel (FFC).

Mature Firms

Source: Authors’ calculations based on data from Banco de México.

Notes: Shaded areas represent 90 and 95 percent confidence intervals. Errors clustered at firm and month level.

Young Firms

Old Firms

Monetary Policy Transmission in an Emerging Market: The Financial- Friction Channel VS The Interest-Rate Channel

Uso Público� �Información de acceso público.

14

4.1 The Financial-Frictions Channel: Firm Age

• We formally test differences and use pairwise comparisons, defining the group of old firms as the reference group.
• The difference in the credit response between young and old firms is statistically significant.
• However, the difference between mature and old firms is not.

Firm level credit response to monetary policy shocks by age

Source: Authors’ calculations based on data from Banco de México.

Notes: Shaded areas represent 90 and 95 percent confidence intervals. Errors clustered at firm and month level.

Difference between Young and Old firms

Difference between Mature and Old firms

Monetary Policy Transmission in an Emerging Market: The Financial- Friction Channel VS The Interest-Rate Channel

Uso Público� �Información de acceso público.

15

4.1 The Financial-Frictions Channel: Firm Size

• The effect also varies across firm-size: for small firms credit growth declines significantly 12 months after the shock and remains depressed for up to 2 years, while for large firms does not respond significantly at any horizon.
• The pairwise comparison indicates that the difference is statistically significant from month t+13 onward.
• Similar to our findings for young firms, small firms are considerably more sensitive to monetary policy.

Large Firms

Firm level credit response to monetary policy shocks by size

Source: Authors’ calculations based on data from Banco de México.

Notes: Shaded areas represent 90 and 95 percent confidence intervals. Errors clustered at firm and month level.

Small Firms

Difference between Small and Large Firms

Uso Público� �Información de acceso público.

16

4.1 The Financial-Frictions Channel: Default History

Non-defaulting Firms

Firm level credit response to monetary policy shocks by default status

Source: Authors’ calculations based on data from Banco de México.

Notes: Shaded areas represent 90 and 95 percent confidence intervals. Errors clustered at firm and month level.

Defaulting Firms

Difference between Defaulting and Non-defaulting Firms

• The effect also varies across firms with different credit default history: credit growth falls for non-defaulting firms but the impact is much earlier and stronger for defaulting firms, being the pairwise comparison statistically significant.
• That is, the evidence indicates the financial friction channel is active, regardless of the type of Friction measure we use.

Uso Público� �Información de acceso público.

17

4.2 The Interest-Rate Channel

• The effect of monetary policy also varies across sector with different durability: credit growth diminishes 16 months after monetary policy contraction in all sectors;
• However, the credit growth reduction starts later in non-durable (16 months) than in durable (10 months) than in construction (14 months), and than in services (11 months).
• The magnitude is also smaller in non-durable: 16 months after the monetary shock, a 1 bp tightening reduces credit growth by 0.08 pp in non-durables, versus 0.17 pp in durables, 0.13 pp in construction, and 0.15 pp in services.
• These results indicate that the interest-rate channel is also active.

Non-durables Manufacturing

Firm Level Credit Response to Monetary Policy Shocks by Sector

Source: Authors’ calculations based on data from Banco de México.

Notes: Shaded areas represent 90 and 95 percent confidence intervals. Errors clustered at firm and month level.

Durables Manufacturing

Construction

Services

Monetary Policy Transmission in an Emerging Market: The Financial- Friction Channel VS The Interest-Rate Channel

Uso Público� �Información de acceso público.

18

4.2 The Interest-Rate Channel

• Pairwise comparisons, using the group of non-durable manufacturing as the reference, show that its response is significantly weaker than in other sectors.
• The fact the the reduction in credit growth is later and weaker for non-durables provides support for the other three sectors, providing further support for the interest-rate channel.

Difference between construction and non-durable manufacturing

Firm Level Credit Response to Monetary Policy Shocks by Sector

Source: Authors’ calculations based on data from Banco de México.

Notes: Shaded areas represent 90 and 95 percent confidence intervals. Errors clustered at firm and month level.

Difference between durable and non-durable manufacturing

Difference between services and non-durable manufacturing

Monetary Policy Transmission in an Emerging Market: The Financial- Friction Channel VS The Interest-Rate Channel

Uso Público� �Información de acceso público.

19

Uso Público� �Información de acceso público.

20

4.3. Financial-Frictions Channel VS Interest-Rate Channel

• To assess the relative strength of the 2 channels, we combine the firm age and the sector durability categorizations and end up with 4 groups of firms: (i) young durables; (ii) young non-durables; (ii) non-young durables; and (iv) non-young non-durables.
• Then, we estimate credit growth responses for each of these 4 groups.
• Credit falls earlier and more for young-durables VS non-young non-durables. The two channels are active here: being both young and operating in a durable sector amplifies the response to a tightening monetary policy shock.
• Credit also declines earlier and more for young non-durables: firm age alone (independent of sector) also determines the response.
• Non-young durables exhibit only a modestly stronger response than non-young non-durables.

Young-Non-durables

Firm Level Credit Response to Monetary Policy Shocks by Age-Durability

Source: Authors’ calculations based on data from Banco de México.

Notes: Shaded areas represent 90 and 95 percent confidence intervals. Errors clustered at firm and month level.

Young-Durables

Non-young Durables

Non-young-Non-durables

Monetary Policy Transmission in an Emerging Market: The Financial- Friction Channel VS The Interest-Rate Channel

Uso Público� �Información de acceso público.

21

4.3. Financial Friction Channel VS Interest-Rate Channel

• We perform pairwise comparisons, using the group non-young non-durables as the reference, and find that the difference in credit response between young durables VS non-young non-durables is negative statistically significant.
• Moreover, the difference between young non-durables VS non-young-non-durables is also negative and statistically significant.
• However, we find that the difference is not statistically significant for non-young durable, suggesting that the dimension of durability is less relevant for monetary policy transmission once firm-age is accounted for.
• Overall, while both the FFC and IRC are active, the evidence indicates that the FFC dominates.
• The result run the counter of those observed for advanced economies, where the IRC dominates (Durante et al. 2022).

Difference between young non-durables and non-young non-durables

Firm Level Credit Response to Monetary Policy Shocks by Age-Durability

Source: Authors’ calculations based on data from Banco de México.

Notes: Shaded areas represent 90 and 95 percent confidence intervals. Errors clustered at firm and month level.

Difference between young durables and non-young non-durables

Difference between non-young durables and non-young non-durables

Monetary Policy Transmission in an Emerging Market: The Financial- Friction Channel VS The Interest-Rate Channel

Uso Público� �Información de acceso público.

22

Uso Público� �Información de acceso público.

23

4.4 The Financial-Frictions VS the Interest-Rate Channel: The Role of Bank Capitalization

• We undertake the same exercises and comparisons for 2 groups: the less- and the more-capitalized banks.
• For the less-capitalized banks, a contraction in monetary policy:
• Reduces credit growth earlier and more strongly for young firms, regardless of whether they operate in durable or non-durable sectors: age along makes a difference;
• However, producing durable goods alone does not change this credit response;

Difference between young non-durables and non-young non-durables

Firm Level Credit Response to Monetary Policy Shocks by Age-Durability in Less-capitalized Banks

Source: Authors’ calculations based on data from Banco de México.

Notes: Shaded areas represent 90 and 95 percent confidence intervals. Errors clustered at firm and month level.

Difference between young durables and non-young non-durables

Difference between non-young durables and non-young non-durables

Uso Público� �Información de acceso público.

24

4.5 Bank Capitalization and Employment

• For the more-capitalized banks, credit responses are not significantly different for firms of different age or for operating in sector of different durability.
• The fact that the dominance of the financial friction channel is stronger precisely for the less-capitalized banks provides further support for both the existence and overall dominance of the channel in am EMDE.
• Employment results point in the same direction: we compare the credit growth responses of municipalities more and less exposed to young firms.
• We find that firms in municipalities more exposed to young firms experience larger declines in employment growth after monetary policy tightening.
• However, the evidence for the Interest rate channel is weaker since municipalities more exposed to durable sectors do not necessarily have a different response.
• This, provides again further evidence for the higher strength of the financial friction channel.

Monetary Policy Transmission in an Emerging Market: The Financial- Friction Channel VS The Interest-Rate Channel

Uso Público� �Información de acceso público.

25

4.6 Robustness

• We further undertake the race-horse between the two channels using use two additional measures for financial constraints, firm-size and recent credit delinquency, and find the same qualitative results: both mechanisms are active, but the financial friction channel dominates.
• We address the two concerns regarding the high-frequency shocks: (i) the “information effect,” whereby policy announcements may reveal private information about economic conditions (Nakamura and Steinsson, 2018), and (ii) uncertainty about the central bank’s reaction function, whereby macroeconomic or financial news may alter market expectations of policy (Bauer and Swanson, 2023).
• We further test robustness using an alternative sectoral taxonomy based on the European Classification of Economic Activities (NACE Rev. 2) instead of the baseline NAICS classification.
• Results remain qualitatively similar across all exercises, confirming that our findings are not driven by the specific proxy used to measure financial constraints, information effects, central bank signaling or the sectoral classification adopted.

Monetary Policy Transmission in an Emerging Market: The Financial- Friction Channel VS The Interest-Rate Channel

Uso Público� �Información de acceso público.

26

5. Conclusions

• This paper provides new evidence on the transmission of monetary policy in an EMDE, where the mechanisms through which monetary policy operates remain relatively understudied.
• Using high-frequency monetary policy shocks and firm-level credit microdata, we show that the financial-frictions channel dominates the traditional interest-rate channel in the transmission of monetary policy.
• Firms more exposed to informational and financing frictions exhibit earlier, stronger, and more persistent credit contractions following monetary tightening.
• The results are consistent with theoretical predictions that EMDEs characteristics, such as weaker disclosure standards, greater informality, limited collateral availability, and underdeveloped credit registries, amplify financial frictions and strengthen the financial-frictions channel relative to the interest-rate channel.
• More broadly, the findings suggest that the institutional and financial environment plays a central role in determining which monetary transmission mechanism dominates across economies.

Monetary Policy Transmission in an Emerging Market: The Financial- Friction Channel VS The Interest-Rate Channel

Uso Público� �Información de acceso público.

27

Appendix

Uso Público� �Información de acceso público.

28

2.2 Summary Statistics by Sector Group

• The services sector account for the largest share in the total number of firms with credit (72%):
• Yet service firms receive loans of a smaller amount (9.07), and thus, their share in total outstanding credit is much lower (47%).

Service firms are small and, at the least in the intensive margin, they could be credit-constrained

• The dominance in the number of firms does not extend to young firms: services account for only 46% of young firms, comparable to other sector.
• The number of bank relationships, a measure of how it can be access to credit, is lowest in construction and services (1.34).

Small does not mean young and small does not always mean the most credit-constrained

Summary Statistics by Sector Group

Source: Authors’ calculations based on data from Banco de Mexico.

Notes: Banks are classified as more or less capitalized based on whether their average pre-sample net capital-to-total-assets ratio is above or below the median across all banks. Firm-level data from January 2011 to December 2019. Following Banco de México’s methodology, a firm is classified as small in a given month if it has not obtained a loan exceeding 100 million pesos (in 2018 constant prices). Average loan amounts are in Mexican pesos at constant prices of 2018. 1/ Shares are calculated monthly and then averaged over time. 2/ Statistics are averaged across firms and months. ***, **, and * denote statistical significance at the 1%, 5%, and 10% levels, respectively.

This dominance is even stronger among small firms (73%).

Uso Público� �Información de acceso público.

29

2.2 Summary Statistics by Age Group

• Young firms account for the largest share in the total number of firms (45%).
• However, they receive loans of smaller amounts (8.96 million pesos) and account for a smaller share of total credit (29%).
• Unlike service firms, they clearly have relationships with fewer banks (1.27 on average) compared to mature and old firms (1.42 & 1.5).

Young always means credit-constrained

• Moreover, more-capitalized banks allocate a larger share of their credit to young firms (38%) than less-capitalized banks (27%), consistent with greater-capitalized banks being more willing to lend to borrowers perceived as riskier.

Summary Statistics by Age Group

Source: Authors’ calculations based on data from Banco de Mexico.

Notes: Firms are classified as young (0–10 years), mature (11–20 years), or old (>20 years). Banks are classified as more or less capitalized according to whether their average pre-sample net capital-to-total-assets ratio is above or below the median. Firm-level data from January 2011 to December 2019. Following Banco de México’s methodology, a firm is classified as small if it has not obtained a loan exceeding 100 million pesos (in 2018 constant prices) up to a given month. Loan amounts are in Mexican pesos at constant prices of 2018. 1/ Shares are calculated monthly and averaged over time. 2/ Statistics are averaged across firms and months. ***, **, and * denote statistical significance at the 1%, 5%, and 10% levels, respectively.

This dominance is NOT stronger among small firms.

Uso Público� �Información de acceso público.

30

3.1 Monetary Policy Shocks

Monetary Policy Transmission in an Emerging Market: The Financial- Friction Channel VS The Interest-Rate Channel

Uso Público� �Información de acceso público.

31

3.1 Monetary Policy Shocks

• Surprises are less predictable than policy-rate changes: like most central banks, Banco de Mexico communicates before announcements, so that part (most) of the monetary policy changes are already incorporated in swap prices.
• 2014: the central bank reduced the monetary policy rate by 50 basis points when no changes or lower reductions were expected according to both market analysts and expectations embedded in market prices.
• 2016: in an extraordinary (non-anticipated) meeting, the central bank increased the policy rate by 50 basis points not due to a specific data point publication.

​

​

​

Monetary Policy Shocks and Policy Rate in Mexico

Source: Banco de México and Banco de México.

Monetary Policy Transmission in an Emerging Market: The Financial- Friction Channel VS The Interest-Rate Channel

Uso Público� �Información de acceso público.

32

3.1 Monetary Policy Shocks

• We perform robustness checks on these surprises. The literature has risen concerns these surprises may not accurately represent pure monetary policy shocks in 2 types of situations:
1. “Information effect” (Nakamura and Steinsson, 2018): when the private sector interprets the policy decision not only as a change in the policy stance, but also as a signal about the central bank’s private information regarding the economy.
1. For example, an unexpected increase in interest rates could be interpreted as a signal that, according to the central bank’s information, future growth will be stronger than expected; this may rise expectations and affect the outcome variable itself.
2. Thus, we apply the “poor man’s” adjustment method of Jarociński and Karadi (2020): It uses additional information surprises that are likely to contain information effects: it excludes surprises in which the rate increases but the sock market increases as well.
3. We exclude all situations in which the shock series where swap rates and stock prices move in the same direction.
2. When the private sector is uncertain about how the central bank reacts (the so-called “Taylor rule”), news releases before the announcement may lead to the expectation of a different central bank response than what actually occurs.
1. This gap can affect the surprise series, and the news itself may influence the outcome variable, creating a correlation unrelated to a true shift in policy stance.
2. To address the second concern, we follow Bauer and Swanson (2023) and conduct a first-stage regression of the surprise series on economic and financial news released prior the announcement, purging the series of information to which the central might be responding with an intensity different from what markets anticipated.

Monetary Policy Transmission in an Emerging Market: The Financial- Friction Channel VS The Interest-Rate Channel

Uso Público� �Información de acceso público.

33

4.3. The Financial-Frictions VS the Interest-Rate Channel: Financial-Frictions Measures based on Size

Small Non-durables

Firm Level Credit Response to Monetary Policy Shocks by Size-Durability

Source: Authors’ calculations based on data from Banco de México.

Notes: Shaded areas represent 90 and 95 percent confidence intervals. Errors clustered at firm and month level.

Small Durables

Large Durables

Large Non-durables

Uso Público� �Información de acceso público.

34

4.3. The Financial-Frictions VS the Interest-Rate Channel: Financial-Frictions Measures based on Size

Difference between small non-durables and large non-durables

Firm Level Credit Response to Monetary Policy Shocks by Size-Durability

Source: Authors’ calculations based on data from Banco de México.

Notes: Shaded areas represent 90 and 95 percent confidence intervals. Errors clustered at firm and month level.

Difference between small durables and large non-durables

Difference between large durables and large non-durables

Uso Público� �Información de acceso público.

35

4.3. The Financial-Frictions VS the Interest-Rate Channel: Financial-Frictions Measures based on Default

Defaulting Non-durables

Firm Level Credit Response to Monetary Policy Shocks by Defaulting Status-Durability

Source: Authors’ calculations based on data from Banco de México.

Notes: Shaded areas represent 90 and 95 percent confidence intervals. Errors clustered at firm and month level.

Defaulting Durables

Non-defaulting Durables

Non-defaulting Non-durables

Uso Público� �Información de acceso público.

36

4.3. The Financial-Frictions VS the Interest-Rate Channel: Financial-Frictions Measures based on Default

Difference between defaulting non-durables and non-defaulting non-durables

Firm Level Credit Response to Monetary Policy Shocks by Defaulting Status-Durability

Source: Authors’ calculations based on data from Banco de México.

Notes: Shaded areas represent 90 and 95 percent confidence intervals. Errors clustered at firm and month level.

Difference between defaulting durables and non-defaulting non-durables

Difference between non-defaulting durables and non-defaulting non-durables

Uso Público� �Información de acceso público.

37

4.4. The Financial-Frictions VS the Interest-Rate Channel: The Role of Bank Capitalization

• Financial frictions arise at both the firm and bank levels, as access to external funding depends on sufficient net worth (Gertler and Kiyotaki, 2015; Kashyap and Stein, 2000; Gerali et al., 2010).
• For banks, lower capitalization reduces loss-absorption capacity, limits funding access, and constrains risk-taking (Gertler and Kiyotaki, 2015; Gerali et al., 2010).
• By affecting capitalization (net worth), monetary policy can therefore shape these borrowing constraints (Gerali et al., 2010; Gertler and Kiyotaki, 2015).
• A contractionary monetary policy can worsen repayment conditions, increase delinquency, and reduce asset prices, generating capital losses that lead banks—especially less-capitalized ones—to curtail lending to riskier borrowers (e.g., young firms), consistent with evidence from the risk-taking channel (Borio and Zhu, 2012; Jiménez et al., 2014; Ioannidou et al., 2015).
• To assess this mechanism, we classify banks using their ICAP—net capital over risk-weighted assets—, split the sample into more- and less-capitalized banks based on the system median during the pre-estimation period, and re-estimate local projections within each subsample.

Uso Público� �Información de acceso público.

38

4.4 The Financial-Frictions VS the Interest-Rate Channel: The Role of Bank Capitalization

• Among less-capitalized banks, credit to both young durables and young non-durables declines significantly more following monetary tightening than credit to non-young non-durables, indicating that credit to young firms contracts more strongly regardless of sector durability.
• In contrast, the response of non-young durables is not statistically different from that of non-young non-durables, suggesting that producing durable goods alone does not independently shape credit responses among less-capitalized banks.

Difference between young non-durables and non-young non-durables

Firm Level Credit Response to Monetary Policy Shocks by Age-Durability in Less-capitalized Banks

Source: Authors’ calculations based on data from Banco de México.

Notes: Shaded areas represent 90 and 95 percent confidence intervals. Errors clustered at firm and month level.

Difference between young durables and non-young non-durables

Difference between non-young durables and non-young non-durables

Uso Público� �Información de acceso público.

39

4.4 The Financial-Frictions vs the Interest-Rate Channel: The Role of Bank Capitalization

• Among more-capitalized banks, the credit growth responses of young durables, young non-durables, and non-young durables do not differ significantly from those of non-young non-durables at any horizon.
• Consistent with the literature on the risk-taking channel (Borio and Zhu, 2012; Jiménez et al., 2014; Ioannidou et al., 2015), these results suggest that the financial-frictions transmission mechanism is more relevant among less-capitalized banks.

Difference between young non-durables and non-young non-durables

Firm Level Credit Response to Monetary Policy Shocks by Age-Durability in More-Capitalized Banks

Source: Authors’ calculations based on data from Banco de México.

Notes: Shaded areas represent 90 and 95 percent confidence intervals. Errors clustered at firm and month level.

Difference between young durables and non-young non-durables

Difference between non-young durables and non-young non-durables

Uso Público� �Información de acceso público.

40

Uso Público� �Información de acceso público.

41

4.5 Channels of Transmission to Employment

• Although we lack firm-level information on real outcomes such as sales or production, we use highly disaggregated employment data from the Mexican Social Security Institute (IMSS) to provide initial evidence on the effects of monetary policy on employment growth.
• The IMSS dataset covers all formal private-sector workers in Mexico and provides fine regional disaggregation, allowing us to construct municipality-level employment growth analogously to the credit analysis and estimate local projections using the same empirical framework.
• To assess the financial-frictions mechanism, we classify municipalities according to the composition of credit in their local banking system during the pre-sample period.
• Specifically, municipalities in the top tercile of credit exposure to young firms are classified as “young”; among the remaining municipalities, those in the top tercile of exposure to mature firms are classified as “mature”; all remaining municipalities are classified as “old.”
• To assess the interest-rate channel, we classify municipalities according to the sectoral composition of local credit exposure during the pre-sample period.
• Specifically, municipalities are sequentially classified into non-durables, construction, durables, and services according to the largest share of local credit allocated to firms in each sector during the pre-sample period.

Uso Público� �Información de acceso público.

42

4.5 Channels of Transmission to Employment

• Municipalities with a larger presence of young and mature firms exhibit stronger declines in employment growth following monetary tightening than municipalities dominated by old firms.
• Thus, consistent with the credit analysis, employment responses vary systematically across age groups, providing further evidence in support of the financial-frictions transmission mechanism.

Municipality level employment response to monetary policy shocks by age

Source: Authors’ calculations based on data from Banco de México.

Notes: Shaded areas represent 90 and 95 percent confidence intervals. Errors clustered at firm and month level.

Difference between Young and Old Municipalities

Difference between Mature and Old Municipalities

Uso Público� �Información de acceso público.

43

4.5 Channels of Transmission to Employment

• Employment responses in non-durable manufacturing are statistically weaker than in services, weaker than in construction—with the difference only close to statistical significance—and not statistically different from durable manufacturing.
• Overall, the results provide some evidence consistent with the interest-rate channel for employment, although the evidence remains relatively weak.

Difference between construction and non-durables manufacturing

Municipality Level Employment Response to Monetary Policy Shocks by Sector

Source: Authors’ calculations based on data from Banco de México.

Notes: Shaded areas represent 90 and 95 percent confidence intervals. Errors clustered at firm and month level.

Difference between durables manufacturing and non-durables manufacturing

Difference between services and non-durables manufacturing

Uso Público� �Información de acceso público.

44

4.5 Channels of Transmission to Employment

• To compare the relative strength of the financial-frictions and interest-rate channels, we combine municipalities’ age and sectoral durability classifications, following the same approach as in the credit analysis.
• Employment growth responses in young non-durables municipalities are statistically significantly stronger than in non-young non-durables municipalities, whereas the responses of non-young durables and non-young non-durables municipalities do not differ significantly.
• Overall, these results suggest that the financial-frictions channel dominates the interest-rate channel also in the transmission of monetary policy to employment growth.

Difference between young non-durables and non-young non-durables

Municipality Level Employment Response to Monetary Policy Shocks by Age-Durability

Source: Authors’ calculations based on data from Banco de México.

Notes: Shaded areas represent 90 and 95 percent confidence intervals. Errors clustered at firm and month level.

Difference between young durables and non-young non-durables

Difference between non-young durables and non-young non-durables

Uso Público� �Información de acceso público.

45