## _wp0651 - Appendix I.

## Source details

**Canonical URL:** [_wp0651 - Appendix I.](https://www.imf.org/-/media/websites/imf/imported-full-text-pdf/external/pubs/ft/wp/2006/_wp0651.pdf)

## Other formats

- [Markdown version](/-/media/websites/imf/imported-full-text-pdf/external/pubs/ft/wp/2006/_wp0651.pdf.md)
- [Structured JSON version](/-/media/websites/imf/imported-full-text-pdf/external/pubs/ft/wp/2006/_wp0651.pdf.json)

---

### I. Introduction and Hypotheses
- Capital controls were widely used until the late 1970s; relaxation in the 1980s increased economic integration. Malaysia’s September 1998 reimposition of controls is central to the debate.
- Two main views motivating firm-level tests:
  - Macroeconomic view: controls as a tool to stabilize in severe external crises.
  - Institutional view: controls as part of relationship-based capitalism that channels lending to approved firms; relaxing controls can invite overborrowing and collapse.
- Firm-level testable implications around crises:
  - Firms with stronger political connections should suffer more when macro shocks reduce government ability to provide advantages.
  - Firms with stronger political connections should benefit more when imposition of capital controls allows increased support for favored firms.
- Key firm-level empirical highlights:
  - In the initial phase (July 1997 to August 1998), roughly 9 percent of the estimated $60 billion loss in market value for politically connected firms may be attributed to the fall in the expected value of their connections.
  - With capital controls in September 1998, up to 32 percent of the estimated $5 billion gain in market value for firms connected to the prime minister may be attributed to the increase in the value of their connections.
  - For connected firms, the value of political connections was estimated in the range of 12–23 percent of their total market value at the end of September 1998.

### II. Brief Chronology of Capital Controls and Policy Responses
- Historical moves:
  - 1968: removal of restrictions on current international transactions; acceptance of IMF Article VIII obligations.
  - 1973: move from fixed to floating exchange rate.
  - 1986–87: gradual liberalization; temporary reintroduction of some controls in 1994.
- Crisis dynamics and responses:
  - After Thailand’s devaluation in July 1997, the ringgit came under severe pressure; offshore ringgit interest rates rose; reserves plummeted.
  - Initial tightening in 1997: spending cuts, 1998 budget target surplus of 2½ percent of GDP; base lending rates rose; later fiscal policy became more expansionary and 1998 budget surplus target relaxed to ½ percent of GDP in March 1998.
  - Early September 1998: capital controls imposed and ringgit pegged to the U.S. dollar. Measures included repatriation requirements for offshore ringgit, prohibition on licensed offshore banks trading ringgit assets, restrictions on ringgit credit between residents and nonresidents, a one-year holding period on nonresidents repatriating proceeds from Malaysian securities sales, and prior approval requirements above certain limits for residents transferring capital abroad.
- Design intention:
  - Controls targeted known offshore supply channels and major portfolio outflows while attempting not to affect foreign direct investment and current account convertibility; presented as temporary.

### III. Macroeconomic Issues: Motivation and Impact
- Official objectives (press releases): (i) limit contagion; (ii) preserve gains from policy measures to stabilize domestic economy; (iii) ensure domestic price and exchange rate stability to revive confidence.
- By summer 1998:
  - A significant portion of capital had already flowed out by imposition date.
  - Ringgit had depreciated by 70 percent and pressure was letting up by summer 1998.
  - Offshore swap differentials were trending down.
  - Quarterly GDP growth showed the crisis had bottomed out in Q1 1998.
- Political/social stability concerns and worst-case domestic capital flight scenario influenced decision to impose controls; worst-case did not materialize.
- Macroeconomic assessment:
  - Evidence does not suggest capital controls made a visible difference in recovery; Malaysia had more favorable initial conditions and stronger institutional indicators relative to some peers.
  - Timing and magnitude of output decline and recovery were similar across Indonesia, the Republic of Korea, Malaysia, and Thailand.
  - Policy choices in Malaysia were broadly orthodox and similar to other crisis countries; fiscal impulse in Malaysia was smaller in 1998 and broadly similar in 1999 compared to peers.
  - Investor ratings fell sharply in 1998–1999 but relative rankings were regained by 2003.
  - FDI “potential” ranking remained around 33rd–34th; actual capital attracted slipped from around 5th–10th pre-crisis to 70th–75th post-crisis (UNCTAD measures).
  - Private fixed investment recovery in Malaysia was slower; private investment as percent of GDP remained under 10 percent and about half the regional average.

### IV. Firm-Level Evidence: Identification and Sample
- Political connections in Malaysia:
  - Two forms: official status for Bumiputra firms and informal ties between politicians and firms run by Malay and Chinese businesspeople.
  - Two influential officials: the Prime Minister (Dr. Mahathir Mohammad; firms denoted PMC) and the Finance Minister (Anwar Ibrahim during the crisis; firms denoted FMC).
- Identification: based on Gomez and Jomo (1997) list; limitations include non-exhaustiveness, potential time-varying connections, unofficial ties, and no measure of connection strength.
- Sample: Malaysian firms from Worldscope; total firms with pre-crisis data = 424; identifiable politically connected = 67; PM connected = 53; FM connected = 14.
- Descriptive patterns (selected exact figures from summary statistics):
  - Stock returns, July 1997–August 1998: All = -78.5%; Politically connected = -83.0%; Unconnected = -77.7%; (p-value) (0.010).
  - Stock returns, September 1998: All = 39.7%; Politically connected = 53.2%; Unconnected = 37.1%; (p-value) (0.000).
  - Size (total assets in $000, pre-crisis): All = 986,606; Politically connected = 1,845,217; Unconnected = 820,423; (p-value) (0.012).
  - Profitability (return on assets, pre-crisis): All = 4.0%; Politically connected = -1.2%; Unconnected = 4.9%; (p-value) (0.041).
  - Leverage (total debt/total assets, pre-crisis): All = 23.7%; Politically connected = 33.7%; Unconnected = 21.9%; (p-value) (0.000).
  - Pre-crisis increase in leverage (one-year): All = 2.7%; Politically connected = 6.3%; Unconnected = 2.0%; (p-value) (0.062).
  - Maturity (short-term debt/total debt): All = 61.8%; Politically connected = 57.1%; Unconnected = 62.8%; (p-value) (0.216).

### V. Crisis-Period Stock-Return Regressions (July 1997–August 1998)
- Regression specification: Stock return_i = α + Political connection_i + Size_i + Leverage_i + Industry_i + ε_i (dividend-inclusive returns in ringgit; size = log total assets; debt ratio = total debt/total assets).
- Table 3 key coefficient estimates (with heteroskedasticity-robust t-statistics):
  - Politically connected: Non-financial firms = -0.075*** [-2.97]; Financial firms = -0.077*** [-3.42]; All firms = -0.077*** [-3.88].
  - PM connected: Non-financial = -0.079*** [-2.78]; Financial = -0.091*** [-3.58]; All = -0.083*** [-3.64].
  - FM connected: Non-financial = -0.059 [-1.61]; Financial = -0.046 [-1.34]; All = -0.056** [-2.06].
  - Firm size: Non-financial = 0.074*** [5.19]; Financial = 0.041* [1.71]; All = 0.070*** [5.56].
  - Debt ratio: Non-financial = -0.0014* [-1.87]; Financial = -0.0011 [-1.65]; All = -0.0014** [-2.10].
- Interpretation:
  - Political connections associated with significantly worse stock performance during the early crisis, consistent with markets pricing a fall in expected government support.
  - Effects robust to inclusion of industry dummies.

### VI. Stock-Return Effects Following Imposition of Controls (September 1998)
- Context: September 1998 combined capital controls announcement (September 1) and ringgit-dollar peg (set early afternoon September 2); Finance Minister removed September 2 and jailed September 20.
- Table 4 key coefficient estimates (September 1998 returns; t-statistics):
  - Politically connected: Non-financial = 0.081 [1.23]; Financial = 0.285 [2.69]; All = 0.138** [2.42].
  - PM connected: Non-financial = 0.130* [1.76]; Financial = 0.403*** [3.02]; All = 0.199*** [2.98].
  - FM connected: Non-financial = -0.116 [-1.11]; Financial = 0.027 [0.24]; All = -0.063 [-0.81].
  - Debt ratio: Non-financial = 0.0036*** [3.48]; Financial = 0.0018 [0.89]; All = 0.0032*** [3.53].
- Cross-sample patterns:
  - Connected financial firms on average had higher September returns of 28.5 percentage points (significant at the 1 percent level as reported in narrative).
  - All firms combined: political connections coefficient shows a higher return of 13.8 percentage points, significant at the 5 percent level.
  - Non-financial PMC firms: higher returns of 13 percentage points, significant at the 10 percent level.
  - Non-financial FMC firms: dummy is minus 11.6 percentage points (not statistically significant).
  - Net difference between PMC and FMC (non-financial) = 24.6 percentage points.
- Interpretation:
  - Market perceived September 1998 events and controls as increasing the value of PM connections while diminishing value of Finance Minister connections.

### VII. Economic Significance: Market-Value Estimates and Explained Variation
- Market-value estimates:
  - Estimated market-value loss during crisis attributable to connections: roughly $5.7 billion of total market value lost by connected firms.
  - Incremental market-value gain for connected firms when controls imposed in September 1998: roughly $1.3 billion.
  - Regression-based estimate: PM-connections account for about a 20 percent increase in firm value in September 1998; in valuations at end-September 1998 this increase would be 12 percent of firm value.
  - Perceived percentage of firm value attributable to connections suggested within the 12 percent to 23 percent range.
- Contribution to explained variation:
  - September 1998 regressions: R-squared rises from 0.109 to 0.143 when political connection variables added — incremental rise of 0.034 (roughly 3.4 percent of total variation).
  - Initial crisis period regressions: R-squared rises from 0.210 to 0.237 when political connection variables added — incremental rise of 0.027 (roughly 2.7 percent of total variation).
- Subsample caveat:
  - In the IFC investable index subset (~109 firms), Minister connections coefficient falls to 0.129 with t-statistic 1.1 (loss of significance noted).

### VIII. Firm-Level Operating Performance: Investment, Growth, Profitability, Leverage (1990–2003)
- Definitions:
  - Investment = capital expenditures / gross fixed assets.
  - Growth = log annual growth rate in sales.
  - Profitability = return on assets.
  - Leverage = total debt / total assets.
- Summary median patterns (Table 5 narrative):
  - Pre-crisis: PMC firms showed higher investment, higher growth, higher leverage, and lower profitability compared with non-PMC firms.
  - Post-crisis: differences in investment and growth largely reversed — PMC firms have lower investment and growth and appear to have even higher leverage than other firms.
- Panel regressions (Table 6 Panel A): Performance_it = α + Firm_i + PMC_i×Crisis_t + PMC_i×Post-crisis_t + Year_t + ε_it (Crisis_t = 1 for 1997–1998; Post-crisis_t = 1 for 1999–2003).
  - Selected coefficients (heteroskedasticity-robust t-statistics):
    - PM connected X crisis: Growth = 0.001 [0.02]; Investment = -0.128* [-1.67]; Leverage = 0.232* [1.92]; Profitability = -0.075 [-1.04].
    - PM connected X post-crisis: Growth = -0.050* [-1.93]; Investment = -0.213*** [-2.91]; Leverage = 0.210 [1.46]; Profitability = -0.003 [-0.11].
  - Number of observations (Panel A): Growth 3035; Investment 3557; Leverage 3786; Profitability 3792.
  - R-squared (Panel A): Growth 0.312; Investment 0.196; Leverage 0.538; Profitability 0.176.
- Period-average cross-sectional regressions (Panels B–D):
  - Pre-crisis (Panel B): PM connected associated with Investment = 0.264** [2.50]; Growth = 0.048 [1.27]; Leverage = 0.049 [1.62]; Profitability = -0.009 [-0.74].
  - Crisis (Panel C): PM connected associated with Leverage = 0.287 [2.27]; Profitability = -0.131* [-1.70]; Growth = 0.053 [1.49]; Investment = 0.022 [0.38].
  - Post-crisis (Panel D): PM connected associations largely diminished; Leverage = 0.241 [1.38]; Growth = 0.008 [0.74]; Investment = -0.013 [-0.16]; Profitability = 0.012 [0.33].
- Robustness:
  - Arellano-Bond GMM and balanced-panel checks yield similar qualitative results (balanced sample ≈ 20 percent of full sample; larger standard errors noted).

### IX. Anecdotal Evidence on Post-Controls Government Support
- Reported types of support to favored firms after controls (from public record and media):
  - Bailouts by the state-owned oil company.
  - Advantageous deals directly from the government.
  - Banking-sector consolidation and repeated debt rollovers for some large companies.
- Authors note these examples are anecdotal and direct measurement of extent of benefits is difficult.

### X. Interpretation and Policy Implications
- Macro conclusion:
  - Authors do not find evidence that Malaysia’s September 1998 controls were essential for recovery or structural reforms.
  - Controls were imposed late and after large depreciation and capital outflows, limiting potential macroeconomic benefits; at best, controls played a preventive role against perceived worst-case risks.
- Firm-level conclusion:
  - Stock market interpreted September 1998 events as helping politically connected firms, especially PM-connected firms.
  - Connected firms were expected to lose benefits in the first phase of the crisis and PM-connected firms to gain when controls were imposed.
  - Presence of political connections does not imply connections caused the crisis or that relationship-based capitalism was necessarily suboptimal.
  - Post-crisis financial performance makes it hard to discern the extent to which connected firms actually received lasting special advantages; possible reasons include reforms reducing privileges or ineffective use of privileges by connected firms.

### XI. Appendix I: Coding and Tables (selected exact entries)
- Coding of firms: based on Gomez and Jomo (1997); firms coded as politically connected if text indicates officers or major shareholders closely related to Prime Minister or Finance Minister (examples preserved in source). Detailed coding available from authors upon request.
- Table: Malaysia — Capital Controls, 1992–2004 (selected exact measures and dates), examples preserved verbatim in source:
  - 4/20/1992 — Total borrowing by residents in foreign currency from domestic commercial and merchant banks to finance imports of goods and services was restricted to the equivalent of RM 1 million (previously there were no limits).
  - 8/4/1997 — Controls were imposed on banks to limit outstanding noncommercial-related ringgit offer-side swap transactions ... to $2 million per foreign customer or its equivalent.
  - 8/28/1997 — A ban on short-selling of the listed securities on KLSE was introduced to limit speculative pressures on stock prices and exchange rates.
  - 9/1/1998 — Multiple measures effective September 1, 1998 including: requirement to repatriate all ringgit held offshore by October 1, 1998 (BNM approval thereafter); approval requirement to transfer funds between external accounts; licensed offshore banks prohibited to trade in ringgit assets; limit introduced on exports and imports of ringgit by residents and nonresident travelers; residents prohibited from granting ringgit credit facilities to nonresident corresponding banks and stockbroking companies; residents prohibited from obtaining ringgit credit facilities from nonresidents; all imports and exports required to be settled in foreign currency; all purchases and sales of ringgit facilities can only be transacted through authorized depository institutions; a 12-month waiting period for nonresidents to convert RM proceeds from sale of Malaysian securities held in external accounts (with exceptions); trading in Malaysian shares on Singapore's CLOB OTC market became de facto prohibited due to enforcement requiring KLSE registration; specific limit on exports of foreign currency by residents and up to the amount brought into Malaysia for nonresidents.
  - 2/15/1999 — The 12-month waiting period replaced with a graduated system of exit levy on the repatriation of the principal of capital investments made prior to February 15, 1999: levy 30 percent if repatriated less than 7 months after entry, 20 percent if repatriated in 7-9 months; and 10 percent if 9-12 months. No levy on principal if repatriated after 12 months.
  - 4/1/2004 — Multiple liberalizations including residents allowed to sell forward nonexport foreign currency receivables for ringgit or another foreign currency; residents with permitted foreign currency borrowing allowed to enter into interest rate swaps with onshore licensed banks; resident individuals with funds abroad allowed to maintain non export foreign currency accounts offshore without any limit imposed on overnight balances; resident companies with domestic borrowing allowed to open non export foreign currency accounts with licensed onshore banks in Malaysia with no limit on overnight balances; the amount of export proceeds that residents may retain in foreign currency accounts with licensed onshore banks increased from $1 million–$70 million to $30 million–$70 million; COFE approval required for issuance of ringgit bonds in Malaysia by MDBs and foreign MNCs; limits for banking institutions on loans to nonresidents raised from RM 200,000 to RM 10,000,000; licensed insurers and takaful operators allowed to invest abroad up to 5% of their margins of solvency and total assets, and up to 10% of net asset value in their own investment-linked funds; unit trust management companies allowed specified foreign investment limits.
- Summary statistics and sample sizes:
  - Number of firms: All = 424; Politically connected = 67; Unconnected = 357; PM connected = 53; FM connected = 14; All (Non-Financial firms only) = 312; Politically connected (Non-Financial) = 50; Unconnected (Non-Financial) = 262.
  - Summary sample-level R-squared (sample split): 0.269 (Non-financial firms); 0.095–0.099 (Financial firms); 0.236–0.237 (All firms).
  - Table 3 dependent variable: stock return from July 1997 to August 1998.
  - Table 4 dependent variable: stock return for September 1998 (number of observations: 302 Non-financial; 111 Financial; 413 All firms).

*Source: _wp0651 - Appendix I (content drawn exclusively from the supplied PDF chapter/section).*

### Appendix I. ............................................................................................................

### _wp0651 - Appendix I.

### Figures
- 1. Malaysia: Cumulative and Net Portfolio Flows, 1997–2000.........................................23
- 2. Malaysia: International Reserves and Exchange Rate, 1995–2002 ...............................23
- 3. Malaysia and Thailand: Swap Differentials, May–December 1998 ..............................24
- 4. Selected Asian Countries: Monetary Indicators, 1998–2001 ........................................25
- 5. Selected Asian Countries: Real GDP Growth, 1996–2001 ...........................................26
- 6. Selected Asian Countries: Political Risk Index, 1995–2002 .........................................27
- 7. Governance Indicators in Percentile Rankings, 1998 ....................................................28
- 8. Asian Countries: Fiscal Indicators, 1995–2000 .............................................................29
- 9. Selected Asian Countries: Private Fixed Investment 1990–2004..................................30

### Tables
- 1. Malaysia: Capital Controls, 1992–2004 ........................................................................31
- 2. Summary Statistics of Firm-Level Sample ....................................................................41
- 3. Political Connections and Crisis-Period Stock Returns.................................................42
- 4. Political Connections and Stock Returns Following Imposition of  
         Capital         Controls .............................................................................................................43
- 5. Political Connections and Median Operating Performance...........................................44

*Source: _wp0651 - Appendix I.*

### 6. Political Connections and Operating Performance: Regression Analysis .....................45

### 6. Political Connections and Operating Performance: Regression Analysis

### I. INTRODUCTION
- Capital controls were widely used until the late 1970s; cautious relaxation in the 1980s increased economic integration.
- Post-Asian crisis, the role of capital controls has been reconsidered, with Malaysia’s September 1998 reimposition central to the debate.
- Two main views:
  - Macroeconomic view: controls as a tool to stabilize in severe external crises (Krugman, 1998; contested by Dornbusch, 2001).
  - Institutional view: controls as part of relationship-based capitalism that channels lending to approved firms; relaxing controls can invite overborrowing and collapse (Rajan and Zingales, 1998, 2003).
- Firm-level testable implications around crises:
  - Firms with stronger political connections should suffer more when macro shocks reduce government ability to provide advantages.
  - Firms with stronger political connections should benefit more when imposition of capital controls allows increased support for favored firms.
- Key firm-level empirical findings summarized in the chapter:
  - In the initial phase (July 1997 to August 1998), roughly 9 percent of the estimated $60 billion loss in market value for politically connected firms may be attributed to the fall in the expected value of their connections.
  - With capital controls in September 1998, up to 32 percent of the estimated $5 billion gain in market value for firms connected to the prime minister may be attributed to the increase in the value of their connections.
  - For connected firms, the value of political connections was in the range of 12–23 percent of their total market value at the end of September 1998.
- Relation to literature:
  - Builds on Fisman (2001) and works linking institutions, firm-level governance, and macro outcomes (Johnson et al., 2000; Mitton, 2002; Lemmon and Lins, 2003; Morck, Yeung, and Yu, 2000).

### II. BRIEF CHRONOLOGY OF CAPITAL CONTROLS AND MACROECONOMIC POLICIES
- Historical policy moves:
  - 1968: Malaysia removed restrictions on payments and transfers for current international transactions, accepting IMF Article VIII obligations.
  - 1973: Further relaxations and move from fixed to floating exchange rate.
  - 1986–87: Gradual liberalization of capital controls, with temporary reintroduction of some controls in 1994 to stem inflows of short-term capital.
- Crisis dynamics:
  - After Thailand devalued in July 1997, the Malaysian ringgit came under severe pressure; portfolio outflows intensified and foreign exchange reserves plummeted.
  - Offshore ringgit interest rates rose relative to onshore rates as speculative positions increased in offshore markets.
- Policy responses:
  - Initial tightening in 1997: spending cuts and a 1998 budget target surplus of 2½ percent of GDP; base lending rates allowed to rise; lending targets adjusted.
  - Early 1998: fiscal policy became more expansionary; 1998 budget surplus target relaxed to ½ percent of GDP in March 1998; financial sector strengthening measures introduced.
  - Early September 1998: capital controls imposed and ringgit pegged to the U.S. dollar. Measures included repatriation requirements for offshore ringgit, prohibition on licensed offshore banks trading ringgit assets, restrictions on residents granting/receiving ringgit credit vis-à-vis nonresidents, a one-year holding period on nonresidents repatriating proceeds from Malaysian securities sales, and prior approval requirements above certain limits for residents transferring capital abroad.
- Design and intent of controls:
  - Controls aimed to close known channels and loopholes for ringgit offshore supply and major portfolio outflows while attempting not to affect foreign direct investment and current account convertibility.
  - Controls were presented as temporary; implementation facilitated by history of controls, effective state capacity, and strong bank supervision.

### III. MACROECONOMIC ISSUES
A. Understanding Motivation for Controls
- Official objectives in press releases: (i) limit contagion effects of external developments; (ii) preserve gains from policy measures to stabilize the domestic economy; (iii) ensure stability in domestic prices and the ringgit exchange rate to revive investor and consumer confidence and facilitate recovery.
- By summer 1998, risks to financial stability had diminished:
  - A significant portion of capital had already flowed out by imposition date.
  - Ringgit had depreciated by 70 percent and pressure was letting up by summer 1998.
  - Offshore swap differentials were trending down.
  - Quarterly GDP growth showed the crisis had bottomed out in Q1 1998.
- Political and social stability concerns influenced the decision; perceived political risk indicators worsened in summer 1998.
- Worst-case scenario (domestic capital flight and increased offshore speculation) could have entailed significant costs; controls were seen as guarding against that eventuality.
- The worst-case scenario did not materialize; improvements in market sentiment and global liquidity contributed, and ex post undervaluation of the ringgit reduced incentives to avoid controls.

B. Macroeconomic Impact
- Kaplan and Rodrik (2001) argued capital controls enabled faster, less painful recovery; this is contested:
  - Malaysia had more favorable initial conditions (e.g., lower burden of short-term corporate debt) and stronger institutional indicators (government effectiveness, regulatory quality, rule of law, control of corruption).
  - Timing and magnitude of output decline and recovery were similar across the four major crisis countries (Indonesia, Republic of Korea, Malaysia, Thailand).
  - Policy choices in Malaysia were broadly orthodox and similar to other crisis countries; timing and pace of interest rate reductions were not unusually aggressive post-controls.
  - Fiscal impulse in Malaysia was smaller in 1998 and broadly similar in 1999 compared to peers.
- Evidence does not suggest capital controls made a visible difference in recovery; responsible macroeconomic policies, structural reforms, initial conditions, and institutional capacity were central.
- Capital controls did not leave clear lasting costs on access to international portfolio capital in the medium term: Malaysia’s investor ratings fell sharply in 1998–1999 but relative rankings were regained by 2003.
- FDI performance: Malaysia’s ranking in FDI “potential” remained around 33rd–34th; actual capital attracted slipped from around 5th–10th before the crisis to 70th–75th after the crisis (UNCTAD measures).
- Private fixed investment recovery in Malaysia has been slower than other crisis countries; private investment as percent of GDP remained under 10 percent and about half the regional average.

### IV. FIRM-LEVEL EVIDENCE
A. Political Connections in Malaysia
- Two forms of political connections prior to 1997 (Gomez and Jomo, 1997):
  - Official status for firms run by ethnic Malays (Bumiputras).
  - Informal ties between leading politicians and firms run by both Malay and Chinese businesspeople.
- Bumiputras comprise about 60 percent of population; NEP (New Economic Policy, 1970) provided privileges (priority for contracts, access to capital, privatized assets, subsidies).
- Two influential officials in promoting firms during the 1990s:
  - The Prime Minister (Dr. Mahathir Mohammad, Prime Minister from 1981 to 2003) — firms denoted PMC (prime minister connected).
  - The Finance Minister (Anwar Ibrahim during the Asian crisis) — firms denoted FMC (finance minister connected).

B. Identifying Firm-Level Political Connections
- Political connections identified based on Gomez and Jomo (1997); used in other studies (Johnson and Mitton, 2003; Faccio, 2005; Faccio, Masulis, and McConnell, 2005; Chong, Liu, and Tan, 2005).
- Limitations:
  - Not exhaustive identification of every connected firm.
  - Some connections identified earlier in the 1990s might have disappeared prior to crisis (though government stability suggests this is unlikely to be prevalent).
  - Many connections are unofficial and not independently verified.
  - Coding does not measure strength of connections.
- The authors use Gomez and Jomo’s list as a systematic proxy for investor perceptions of political connections during the period.

C. Sample and Descriptive Statistics
- Sample drawn from Malaysian firms in the Worldscope database; includes active and inactive firms, representative of main board of Kuala Lumpur Stock Exchange. Total firms with available pre-crisis data: 424; identifiable political connections: 67.
- Key descriptive comparisons (pre-crisis and crisis periods; see Table 2 summarized):
  - Row 2: Politically connected firms had significantly worse returns than unconnected firms during crisis period July 1997–August 1998; no significant difference between PMC and FMC in that interval.
  - Row 3: Politically connected firms had significantly better returns after imposition of capital controls in September 1998; PMC firms performed much better than FMC firms in this period.
  - Row 4: No significant differences in returns after September 1998 between politically connected and unconnected firms.
  - Row 5: Politically connected firms were significantly larger in total assets (about twice the size on average) compared with unconnected firms.
  - Row 6: Asset growth immediately before the crisis was not significantly greater in connected firms.
  - Row 7: Politically connected firms appeared less profitable in terms of return on assets before the crisis (but regression controls for other characteristics show no evidence of lower profitability before the crisis).
  - Rows 8–10: No significant differences in liquidity (current ratio), efficiency (asset turnover), or book-to-market ratios before the crisis.
  - Row 11: Connected firms had debt-asset ratios more than 11 percentage points higher, on average, than unconnected firms prior to the crisis.
  - Row 12: Leverage was rising significantly faster for connected firms prior to the crisis.
  - Row 13: Connected firms had less short-term debt (maturity < 1 year) as a percentage of total debt.
  - Row 14: Connected firms had a lower percentage increase in short-term debt prior to the crisis.
- Regression controls for size, leverage, and industry; after controlling for determinants of leverage, connected non-financial firms still had debt ratios five percentage points higher (coefficient significant at the 10 percent level).
- Overall: differences in size and leverage are primary pre-crisis distinctions between connected and unconnected firms; no strong evidence connected firms operated meaningfully better pre-crisis.

D. Hypotheses and Regression Specification
- Rajan and Zingales (1998) motivated hypotheses:
  - The stock price of politically connected firms should have fallen more in the early crisis period.
  - When capital controls were imposed, politically connected firms should have risen relative to unconnected firms; benefits concentrated in PMC rather than FMC firms in September 1998.
  - After controls, PMC firms should show evidence of having received advantages.
- Data and timing:
  - Crisis period defined as July 1997 through August 1998 (monthly stock return data).
  - September 1998 treated as key date for imposition of capital controls and for assessing stock price impact (controls announced September 1; ringgit-dollar peg set early afternoon September 2, 1998).
- Regression specification (cross-sectional return model):
  - Stock return_i = α + Political connection_i + Size_i + Leverage_i + Industry_i + ε_i
  - Stock returns are dividend-inclusive and expressed in ringgit.
  - Size and Leverage defined as in Table 2; Industry corresponds to dummy variables for broad industries (Campbell, 1996).

E. The Crisis Period: July 1997–August 1998 (Regression Findings)
- Table 3 regression results summary:
  - For nonfinancial firms, the politically connected dummy coefficient is -0.075, indicating political connection associated with a greater stock price decline of 7.5 percentage points, on average, during July 1997–August 1998.
  - For financial firms, the coefficient is -0.077.
  - These coefficients are significant at the 1 percent level of confidence.
  - Control variables: larger size associated with higher returns during the crisis; higher leverage associated with lower returns.
  - Using separate dummies for PMC and FMC: both types had worse performance than unconnected firms; PMC decline of 7.9 percentage points, FMC decline of 5.9 percentage points among non-financial firms in this period.
  - Up to six of 12 industry dummies can be significant in crisis-period regressions, but including industry dummies does not weaken political connection coefficients.
- Interpretation:
  - Political connections were associated with significantly worse stock price performance in the first phase of the crisis, consistent with the Rajan and Zingales view that expected government support declined.
  - Market perceived policies of July 1997–August 1998 as squeezing politically connected firms; rhetoric and stated policies emphasized budget discipline and reducing cronyism.

F. Effects of Capital Controls: September 1998
- Conceptual expectation:
  - If connections lost value during early crisis, imposition of capital controls should yield a larger rebound for connected firms than for unconnected firms (market perception of events associated with controls).
- Malaysian-specific identification:
  - September 1998 combined imposition of controls and downfall of the Finance Minister (fired September 2, 1998; jailed on corruption charges September 20, 1998).
  - Market perceived these events as reducing value of connections for firms tied to the Finance Minister; PMC firms thus expected to benefit more than FMC firms in September 1998.
- Empirical lead-in:
  - Table 4 presents regressions of stock returns for September 1998 on the same variables as Table 3.
  - Politically connected firms as a whole enjoyed a rebound in September 1998 (total increase in average stock price: 53.2 percent for connected firms versus 37.1 percent for unconnected firms; see Table 2).

*Source: IMF working paper chapter content (6. Political Connections and Operating Performance: Regression Analysis).*

### 8.1 percentage points, not significant at standard levels, may be attributed to political

### _wp0651 - 8.1 percentage points, not significant at standard levels, may be attributed to political

### Stock-return effects of political connections around the crisis and September 1998 capital controls
- Connected firms among financial firms: on average had a higher return of 28.5 percentage points, significant at the 1 percent level.
- Connected firms for all firms combined: political connections coefficient shows a higher return of 13.8 percentage points, significant at the 5 percent level.
- Non-financial firms: PMC firms on average experienced higher returns of 13 percentage points, significant at the 10 percent level.
- Non-financial FMC firms: dummy is minus 11.6 percentage points (not statistically significant).
- Net difference between PMC and FMC firms (non-financial): 24.6 percentage points (13 plus 11.6).
- Financial firms: PMC firms had higher returns of 40.3 percentage points, significant at the 1 percent level.
- All firms combined: PMC firms on average had higher returns of 19.9 percentage points, significant at the 1 percent level; FMC on average had lower returns of 6.3 percentage points (not statistically significant).
- Cross-sectional test on foreign capital access: politically connected firms without foreign capital access performed better than connected firms with foreign capital access when capital controls were imposed (evidence mixed at times; reference to Johnson and Mitton, 2003).

### Economic significance (market-value estimates and explanatory power)
- Estimated market-value loss during the crisis attributable to connections: roughly $5.7 billion of the total market value lost by connected firms.
- Incremental market-value gain for connected firms when capital controls were imposed in September 1998: roughly $1.3 billion.
- Regression-based estimate of PM-connections’ impact in September 1998: PM-connections account for about a 20 percent increase in firm value in September 1998.
- In terms of valuations at the end of September 1998: increase would be 12 percent of firm value.
- Suggested perceived percentage of firm value attributable to connections: within the 12 percent to 23 percent range estimated by Fisman (2001) for connected firms in Indonesia (authors note 12 percent as a low estimate).
- Contribution of political connection variables to explained variation in returns:
  - September 1998 regressions: R-squared rises from 0.109 to 0.143 when political connection variables are added — an incremental rise of 0.034, suggesting roughly 3.4 percent of the total variation in returns is explained by differences in political connections (alternatively, about 1/4 of the systematic, explainable variation).
  - Initial crisis period regressions: R-squared rises from 0.210 to 0.237 when political connection variables are added — an incremental rise of 0.027, suggesting 2.7 percent of the total variation in returns is explained by differences in political connections.
- Caveat on liquidity subset: when limiting sample to firms in the IFC’s investable index (about 109 firms, ~1/4 of main sample), the Minister connections coefficient falls to 0.129 with a t-statistic of 1.1 (loss of significance noted).

### Firm-level operating performance: 1990–2003 (investment, growth, profitability, leverage)
- Definitions used:
  - Investment = ratio of capital expenditures to gross fixed assets.
  - Growth = log annual growth rate in sales.
  - Profitability = return on assets.
  - Leverage = ratio of total debt to total assets.
- Median firm-level patterns (Table 5 summary):
  - Pre-crisis years: PMC firms showed higher investment, higher growth, higher leverage, and lower profitability (compared with non-PMC firms).
  - Post-crisis years: differences in investment and growth largely reversed — PMC firms have lower investment and growth.
  - Post-crisis: PMC firms appear to have had even higher leverage than other firms compared with pre-crisis years.
  - Overall: effects of being connected to the PM on operating performance were very different after imposition of capital controls compared with prior to the crisis.
- Panel regression framework (Panel A, Table 6):
  - Model: Performance_it = α + Firm_i + PMC_i×Crisis_t + PMC_i×Post-crisis_t + Year_t + ε_it.
  - Crisis_t = 1 for years 1997–1998; Post-crisis_t = 1 for years 1999–2003.
  - Findings: compared with unconnected firms, PMC firms suffered a large drop in relative investment and growth from the pre-crisis to post-crisis period; they also had less growth and higher leverage, relative to unconnected firms, in the crisis period compared with the pre-crisis period.
  - Robustness: similar results for investment, leverage and profitability using Arellano-Bond GMM; balanced panel gives similar results but larger standard errors (balanced sample ~20 percent of full sample).
- Cross-sectional average regressions (Panels B, C, D):
  - Model: Avg.Performance_i = α + PMC_i + Industry_i + ε_i, where Avg.Performance_i is average over pre-crisis, crisis, or post-crisis period.
  - Results: PMC firms had a growth advantage in the pre-crisis period that disappeared after the crisis; connected firms had more leverage and less profitability during the crisis.
  - Consistency: Panels B, C, and D show similar results when using the same set of firms in each period.
  - Financial vs non-financial: regressions run separately for non-financial firms only did not find significant differences.

### Reported anecdotes and types of government support after capital controls
- Three types of reported benefits to favored firms after controls:
  - Bailouts by the state-owned oil company (examples cited in international media).
  - Advantageous deals directly from the government (examples cited).
  - Banking-sector consolidation and repeated debt rollovers for some large companies (examples cited).
- Note: these are anecdotal extracts from the public record; direct measurement of the extent to which connected firms received benefits is difficult.

### Interpretation and implications (conclusion)
- The authors do not find evidence that Malaysia's September 1998 controls were essential for recovery or structural reforms.
- Malaysia’s macroeconomic performance after capital controls was comparable to other countries recovering from the Asian financial crisis; controls were imposed late and after large depreciation and capital outflows, limiting macroeconomic benefits.
- At best, controls played a preventive role in guarding against perceived risks to financial stability, but this role was not tested by observable pressure.
- Firm-level evidence: stock market interpreted September 1998 events as helping politically connected firms (relative to unconnected firms).
  - Connected firms were expected to lose benefits in the first phase of the crisis; PM-connected firms were expected to gain benefits when capital controls were imposed in September 1998.
- Presence of political connections does not imply connections caused the crisis or that relationship-based capitalism was necessarily suboptimal.
- Based on actual financial performance after the crisis, it is hard to discern the extent to which firms actually received special advantages — possibly because reforms reduced advantages or connected firms did not make effective use of privileges.

*Italic: Content drawn exclusively from the supplied PDF chapter/section.*

### APPENDIX I

### APPENDIX I

### I. Coding of Firms
- Definition: Firms coded as “politically connected” are those that Gomez and Jomo (1997) identify as having officers or major shareholders with close relationships with key government officials—primarily the Prime Minister and the Finance Minister (and their allies).
- Coding procedure:
  - The entire text of Gomez and Jomo (1997) was searched for indications of connections and firms were coded accordingly.
  - Examples given in source text:
    - [Firm A] is coded as politically connected, with the Prime Minister as the primary connection, because Gomez and Jomo state [Firm A] is “controlled by [Person X], who is closely linked to [an ally of the Prime Minister] (p. 103)”.
    - [Firm B] is coded as politically connected with its primary connection listed as the Prime Minister, because Gomez and Jomo state “The chairman of [Firm B] was [Person Y] of the [Group J], a close friend of [the] Prime Minister” (p. 59).
    - [Firm C] is coded as politically connected, with the Finance Minister as the primary connection, because Gomez and Jomo state “[Firm C] (in which [Person Z], probably [the Finance Minister’s] closest confidant, has an interest...)” (p. 57).
- Availability: “The detailed coding is available from the authors upon request.”

### Selected Empirical Figures and Indicators (summarized)
- Figure content (titles preserved; data sources noted):
  - Figure 1. Malaysia: Cumulative and Net Portfolio Flows, 1997–2000 (in billions of U.S. dollars). Source: Meesook and others (2001).
  - Figure 2. Malaysia: International Reserves and Exchange Rate, 1995–2002. Source: IMF, International Financial Statistics database. Labels: “Reserves, end of month US$ billions, left scale” and “Exchange Rate, end of month, Ringgit per US$, right scale.”
  - Figure 3. Malaysia and Thailand: Swap Differentials, May–December 1998. Sources: Data provided by the authorities; Consensus Economics Inc., Asia Pacific Consensus Forecasts; and IMF staff estimates.
  - Figure 4. Selected Asian Countries: Monetary Indicators, 1998–2001. Sources: IMF, International Financial Statistics and Asia Pacific Department databases. Panels include Nominal Interest Rates (Percent per annum, end of period), Real Interest Rate (Percent per annum, end of period), and Private Sector Credit Growth (12-month percent change).
  - Figure 5. Selected Asian Countries: Real GDP Growth, 1996–2001. Sources: IMF, Asia Pacific Department Database for Indonesia, the Republic of Korea, and Thailand, Haver Analytics for Malaysia, and IMF, International Financial Statistics database. Note: “1/ Annualized; seasonally adjusted.”
  - Figure 6. Selected Asian Countries: Political Risk Index, 1995–2002. Source: International Country Risk Guide, available on the Web at www.ICRGonline.com.
  - Figure 7. Governance Indicators in Percentile Rankings, 1998. Source: Kaufmann, Kraay and Mastruzzi (2005). Note: upper-middle-income country group definition provided in source.
  - Figure 8. Asian Countries: Fiscal Indicators, 1995–2000 (in percent of GDP). Source: Meesook and others (2001). Panels include Overall Fiscal Balance and Cumulative Fiscal Impulse.
  - Figure 9. Selected Asian Countries: Private Fixed Investment, 1990–2004 (private gross fixed capital formation as percent of GDP). Note: Observations for Indonesia before 1999 are estimated on an assumption described in source. Sources: IMF, World Economic Outlook Database; OECD database for the Republic of Korea.
- No numeric values beyond figure titles and axes labels are reproduced here because raw figure datapoints are graphical in the source.

### Table: Malaysia — Capital Controls, 1992–2004 (selected measures and dates)
- Overview: Chronological listing of measures affecting borrowing, international transactions in ringgit, inflows/outflows of portfolio and other capital, bank and foreign exchange transactions, stock market transactions, derivatives, and foreign direct investment.
- Selected entries (date — measure — category) preserving exact wording and numeric limits as in source:
  - 1991 — No changes.
  - 4/20/1992 — Total borrowing by residents in foreign currency from domestic commercial and merchant banks to finance imports of goods and services was restricted to the equivalent of RM 1 million (previously there were no limits). — Borrowing in foreign currency domestically and abroad.
  - 7/9/1992 — Borrowing under the Export Credit Refinance Facilities ... by non-resident controlled companies would be considered domestic borrowing. — Borrowing in foreign currency domestically and abroad.
  - 11/1/1992 — Guidelines on foreign equity capital ownership liberalized; companies exporting at least 80 percent of their production were no longer subject to any equity requirements; companies exporting between 50 percent and 79 percent permitted to hold 100 percent equity, provided that they have invested $50 million or more in fixed assets or completed projects with at least 50 percent local value added ... — Foreign direct investment.
  - 1/17/1994 — A ceiling was placed on the net external liability position of domestic banks (excluding trade-related and direct investment inflows) (removed on January 20, 1995). — Bank and foreign exchange transactions.
  - 2/23/1994 — Prohibition of forward transactions (on bid side) and nontrade-related swaps by commercial banks with foreign customers ... (lifted on August 16, 1994). — Bank and foreign exchange transactions.
  - 8/4/1997 — Controls were imposed on banks to limit outstanding noncommercial-related ringgit offer-side swap transactions ... to $2 million per foreign customer or its equivalent. Hedging requirements of foreigners for trade-related and genuine portfolio and foreign direct investment investments were excluded. — Bank and foreign exchange transactions.
  - 8/28/1997 — A ban on short-selling of the listed securities on KLSE was introduced to limit speculative pressures on stock prices and exchange rates. — Stock market transactions.
  - 9/1/1998 — Multiple measures effective September 1, 1998 including: requirement to repatriate all ringgit held offshore by October 1, 1998 (BNM approval thereafter); approval requirement to transfer funds between external accounts; licensed offshore banks prohibited to trade in ringgit assets; limit introduced on exports and imports of ringgit by residents and nonresident travelers; residents prohibited from granting ringgit credit facilities to nonresident corresponding banks and stockbroking companies; residents prohibited from obtaining ringgit credit facilities from nonresidents; all imports and exports required to be settled in foreign currency; all purchases and sales of ringgit facilities can only be transacted through authorized depository institutions; a 12-month waiting period for nonresidents to convert RM proceeds from sale of Malaysian securities held in external accounts (with exceptions); trading in Malaysian shares on Singapore's CLOB OTC market became de facto prohibited due to enforcement requiring KLSE registration; specific limit on exports of foreign currency by residents and up to the amount brought into Malaysia for nonresidents. — International transactions in ringgit; Outflows of portfolio and other capital; Stock market transactions; Export and import of currency.
  - 2/15/1999 — The 12-month waiting period replaced with a graduated system of exit levy on the repatriation of the principal of capital investments made prior to February 15, 1999: levy 30 percent if repatriated less than 7 months after entry, 20 percent if repatriated in 7-9 months; and 10 percent if 9-12 months. No levy on principal if repatriated after 12 months. — Outflows of portfolio and other capital.
  - 3/14/2000 — Funds arising from sale of securities purchased by nonresidents on the CLOB can be repatriated without payment of exit levy. — Outflows of portfolio and other capital.
  - 6/1/2001 — All controls on the trading of futures and options by nonresidents on the MDEX were eliminated. The Commodity and Monetary Exchange of Malaysia and the KLSE were merged to form the MDEX. — Derivatives.
  - 2/1/2001 — The exit levy on profits repatriated after one year from the month the profits are realized was abolished. Portfolio profits repatriated within one year remained subject to the 10 percent levy. — Outflows of portfolio and other capital.
  - 5/1/2001 — The 10 percent exit levy imposed on profits arising from portfolio investments repatriated within one year of realization was abolished. — Outflows of portfolio and other capital.
  - 12/1/2000 — Foreign-owned banks in Malaysia were allowed to extend up to 50 percent (previously 40 percent) of the total domestic credit facilities to nonresident controlled companies, in case of credit facilities extended by resident banks. — Domestic lending by foreign-owned banks.
  - 4/1/2003 — Exporters were allowed to retain a portion of their export proceeds in foreign currency accounts with onshore licensed banks in Malaysia with overnight limits ranging between the equivalent of US$ 1 million and US$ 70 million, or any other amount that has been approved (previously, the limit was between US$ 1 million and US$ 10 million). — International transactions in ringgit; Export proceeds.
  - 4/1/2004 — Multiple liberalizations including residents allowed to sell forward nonexport foreign currency receivables for ringgit or another foreign currency; residents with permitted foreign currency borrowing allowed to enter into interest rate swaps with onshore licensed banks; resident individuals with funds abroad allowed to maintain non export foreign currency accounts offshore without any limit imposed on overnight balances; resident companies with domestic borrowing allowed to open non export foreign currency accounts with licensed onshore banks in Malaysia with no limit on overnight balances; the amount of export proceeds that residents may retain in foreign currency accounts with licensed onshore banks increased from $1 million–$70 million to $30 million–$70 million; COFE approval required for issuance of ringgit bonds in Malaysia by MDBs and foreign MNCs; limits for banking institutions on loans to nonresidents raised from RM 200,000 to RM 10,000,000; licensed insurers and takaful operators allowed to invest abroad up to 5% of their margins of solvency and total assets, and up to 10% of net asset value in their own investment-linked funds; unit trust management companies allowed specified foreign investment limits. — Categories include Outflows of portfolio and other capital; International transactions in ringgit; Bank and foreign exchange transactions; Export proceeds; Foreign direct investment.
- Sources for table: IMF, Annual Report on Exchange Arrangements and Exchange Restrictions; Bank Negara Malaysia, Annual Report and Exchange Notices, various years.

### Summary Statistics: Malaysian Firms (Worldscope sample)
- Sample sizes:
  - Number of firms: All = 424; Politically connected = 67; Unconnected = 357; PM connected = 53; FM connected = 14; All (Non-Financial firms only) = 312; Politically connected (Non-Financial) = 50; Unconnected (Non-Financial) = 262.
- Stock returns:
  - July 1997 to August 1998: All = -78.5%; Politically connected = -83.0%; Unconnected = -77.7%; (p-value) (0.010).
  - September 1998: All = 39.7%; Politically connected = 53.2%; Unconnected = 37.1%; (p-value) (0.000).
  - October 1998 to September 2000: All = 81.9%; Politically connected = 83.5%; Unconnected = 81.7%; (p-value) (0.897).
  - PM connected and FM connected comparisons are reported with p-values in the source table (examples: PM connected September 1998 = 61.7%; FM connected September 1998 = 31.3%; (p-value) (0.021)).
- Pre-crisis performance measures (last reported financial statements prior to July):
  - Size (total assets in $000): All = 986,606; Politically connected = 1,845,217; Unconnected = 820,423; (p-value) (0.012). Non-financial firms: All = 599,554; Politically connected = 1,299,733; Unconnected = 465,535; (p-value) (0.000).
  - Growth (in assets, one-year): All = 50.3%; Politically connected = 67.3%; Unconnected = 46.8%; (p-value) (0.301).
  - Profitability (return on assets): All = 4.0%; Politically connected = -1.2%; Unconnected = 4.9%; (p-value) (0.041).
  - Liquidity (current ratio): All = 1.77; Politically connected = 1.53; Unconnected = 1.82; (p-value) (0.432).
  - Efficiency (asset turnover): All = 0.55; Politically connected = 0.47; Unconnected = 0.56; (p-value) (0.147).
  - Valuation (Book/market ratio): All = 0.45; Politically connected = 0.47; Unconnected = 0.45; (p-value) (0.568).
- Pre-crisis leverage:
  - Leverage (total debt/total assets): All = 23.7%; Politically connected = 33.7%; Unconnected = 21.9%; (p-value) (0.000). Non-financial firms: All = 26.1%; Politically connected = 36.9%; Unconnected = 24.0%; (p-value) (0.000).
  - Pre-crisis increase in leverage (one-year): All = 2.7%; Politically connected = 6.3%; Unconnected = 2.0%; (p-value) (0.062).
  - Maturity (short-term debt/total debt): All = 61.8%; Politically connected = 57.1%; Unconnected = 62.8%; (p-value) (0.216).
  - Pre-crisis increase (one-year): All = -2.2%; Politically connected = -7.7%; Unconnected = -1.1%; (p-value) (0.088).
- Notes: The table presents summary statistics of Malaysian firms in the Worldscope database. The numbers reported are simple averages except as noted. Listed p-values are from t-tests of differences of means. "Politically connected" refers to a firm with identifiable political connections from Gomez and Jomo (1997). A financial firm is defined as one with primary SIC in the range 6000-6999. Financial figures are based on the last reported financial statements prior to July.

*Source: _wp0651 - APPENDIX I, _wp0651 - APPENDIX I (PDF).*

### 1997.  Data points are missing for some items, thus the number of observations included for each average may vary

### _wp0651 - 1997. Data points are missing for some items, thus the number of observations included for each average may vary

### Summary statistics and sample size
- Number of observations: 312 (Non-financial firms), 112 (Financial firms), 424 (All firms).
- R-squared (sample split): 0.269 (Non-financial firms), 0.095–0.099 (Financial firms), 0.236–0.237 (All firms).
- Dependent variable for Table 3 regressions: stock return from July 1997 to August 1998.

### Table 3 — Political Connections and Crisis-Period Stock Returns (July 1997 to August 1998)
- Coefficient estimates (Non-financial firms / Financial firms / All firms) with heteroskedasticity-robust t-statistics in brackets:
  - Politically connected: -0.075*** [-2.97] / -0.077*** [-3.42] / -0.077*** [-3.88]
  - PM connected: -0.079*** [-2.78] / -0.091*** [-3.58] / -0.083*** [-3.64]
  - FM connected: -0.059 [-1.61] / -0.046 [-1.34] / -0.056** [-2.06]
  - Firm size: 0.074*** [5.19] / 0.041* [1.71] / 0.070*** [5.56]
    - repeated entry: 0.074*** [5.19] / 0.042* [1.75] / 0.070*** [5.56]
  - Debt ratio: -0.0014* [-1.87] / -0.0011 [-1.65] / -0.0014** [-2.10]
    - repeated entry: -0.0014* [-1.85] / -0.0010 [-1.53] / -0.0014** [-2.07]

- Note: "Politically connected" from Gomez and Jomo (1997). Firm size = log of total assets; debt ratio = total debt over total assets. Asterisks denote significance: *** 1% level, ** 5% level, * 10% level.

### Table 4 — Political Connections and Stock Returns Following Imposition of Capital Controls (September 1998)
- Dependent variable: stock return for September 1998.
- Number of observations: 302 (Non-financial firms), 111 (Financial firms), 413 (All firms).
- R-squared: 0.142 / 0.115 / 0.128 (first specification), 0.154 / 0.153 / 0.143 (second specification).
- Coefficient estimates (Non-financial firms / Financial firms / All firms) with t-statistics:
  - Politically connected: 0.081 [1.23] / 0.285 [2.69] / 0.138** [2.42]
  - PM connected: 0.130* [1.76] / 0.403*** [3.02] / 0.199*** [2.98]
  - FM connected: -0.116 [-1.11] / 0.027 [0.24] / -0.063 [-0.81]
  - Firm size: 0.014 [0.42] / -0.038 [-0.50] / 0.001 [0.04]
    - alternate: 0.015 [0.43] / -0.043 [-0.58] / 0.000 [0.01]
  - Debt ratio: 0.0036*** [3.48] / 0.0018 [0.89] / 0.0032*** [3.53]
    - alternate: 0.0035*** [3.40] / 0.0012 [0.58] / 0.0031*** [3.35]

- Interpretation evident in coefficients: politically connected and PM connected firms show positive coefficients for September 1998 returns, with statistical significance in several specifications. Debt ratio shows a positive and highly significant coefficient in non-financial and all-firms specifications.

### Table 5 — Political Connections and Median Operating Performance (1990–2003)
- Table reports median operating performance for Malaysian firms for the years 1990 to 2003. Comparative investment figures are given for Thailand and Korea.
- Performance measures defined:
  - Investment = capital expenditures / gross fixed assets
  - Growth = log annual real growth rate in sales
  - Leverage = total debt / total assets
  - Profitability = return on assets
- "PM effect" refers to coefficient estimates for a "PM connected" indicator from regressions of performance measures on a PM indicator for each year 1990 to 2003. A full set of 2-digit SIC dummies and a control for firm size are included but not reported.
- Annual median values listed for 1990–2003 for Malaysia, with comparative columns for Thailand and Rep. of Korea (original table presents medians by year; exact annual median numbers appear in the table).

### Table 6 — Political Connections and Operating Performance: Regression Analysis (Panel and Period Averages)
- Panel A: 1990–2003 (panel regressions with firm-fixed effects and year dummies). Numbers in brackets are heteroskedasticity-robust t-statistics adjusted for firm-level clustering.
  - PM connected X crisis: 0.001 [0.02] (Growth) ; -0.128* [-1.67] (Investment) ; 0.232* [1.92] (Leverage) ; -0.075 [-1.04] (Profitability)
  - PM connected X post-crisis: -0.050* [-1.93] (Growth) ; -0.213*** [-2.91] (Investment) ; 0.210 [1.46] (Leverage) ; -0.003 [-0.11] (Profitability)
  - Number of observations (Panel A): Growth 3035; Investment 3557; Leverage 3786; Profitability 3792
  - R-squared (Panel A): 0.312 (Growth), 0.196 (Investment), 0.538 (Leverage), 0.176 (Profitability)

- Panels B, C, D: regress average performance measures over pre-crisis, crisis, and post-crisis periods on political connection indicator (industry dummies included).
  - Selected PM connected coefficients and statistics across panels (examples reported in table):
    - Panel B / Pre-crisis: PM connected 0.048 [1.27] (Growth) ; 0.264** [2.50] (Investment) ; 0.049 [1.62] (Leverage) ; -0.009 [-0.74] (Profitability); N = 279–324; R-squared varies by measure.
    - Panel C / Crisis: PM connected 0.053 [1.49] (Growth) ; 0.022 [0.38] (Investment) ; 0.287 [2.27] (Leverage) ; -0.131* [-1.70] (Profitability); N = 283–355; R-squared varies by measure.
    - Panel D / Post-crisis: PM connected 0.008 [0.74] (Growth) ; -0.013 [-0.16] (Investment) ; 0.241 [1.38] (Leverage) ; 0.012 [0.33] (Profitability); N = 287–355; R-squared varies by measure.

- Notes on interpretation:
  - Pre-crisis, crisis, and post-crisis periods defined as 1990–1996, 1997–1998, and 1999–2003 respectively.
  - Asterisks denote levels of significance: *** 1% level, ** 5% level, * 10% level.

### Key empirical patterns (as reported in tables)
- During the Asian crisis (July 1997–August 1998), political connections and PM connections are associated with statistically significant negative coefficients on crisis-period stock returns (Table 3).
- Following imposition of capital controls (September 1998), politically connected firms and PM-connected firms show positive and in some cases statistically significant stock return coefficients (Table 4).
- In panel regressions of operating performance (1990–2003), PM-connected firms exhibit time-varying associations:
  - PM connected X crisis: negative association with Investment (−0.128*), positive association with Leverage (0.232*).
  - PM connected X post-crisis: negative association with Growth (−0.050*), negative association with Investment (−0.213***), positive association with Leverage (0.210).
- Across period-average regressions, PM connections are associated with:
  - Pre-crisis: positive and significant association with Investment (0.264**).
  - Crisis: positive and significant association with Leverage (0.287) and negative association with Profitability (−0.131*).
  - Post-crisis: limited statistically significant positive associations for Leverage (0.241) in reported specification.

*Source: _wp0651 - 1997. Data points are missing for some items, thus the number of observations included for each average may vary*

---


_Source: https://www.imf.org/-/media/websites/imf/imported-full-text-pdf/external/pubs/ft/wp/2006/_wp0651.pdf_