The Politics of Corruption

Lecture 5: The Costs of Corruption

Bogdan G. Popescu
bgpopescu@tec.mx

Tecnológico de Monterrey

Last time: Sofia paid $40 — and it worked

  • At the border, $20 made the barrier lift and her cargo move.
  • At the permit office, another $20 got the permit the same day.
  • Without those shortcuts, the delays would have cost her about $500.
  • She paid $40 to avoid roughly $500 in losses.

For Sofia, paying the bribes was rational — and made her business more efficient.

Sofia with her shipping papers after paying to get her cargo moving, with the next administrative step ahead.

Sofia after the barrier lifts, continuing through the same corrupt system.

Then where is the cost?

Sofia’s private calculation

  • Delay avoided: about $500
  • Bribes paid: $40
  • Immediate gain from paying: about $460

The two officials gain the $40 she pays. The money itself mostly changes pockets.

The system-level question

  • Why was Sofia facing the $500 cost in the first place?
  • Would the $500 cost still exist if officials could not collect bribes?
  • What happens to a trader who cannot afford the $40 shortcut?
  • What happens if thousands of firms face the same choice?

A bribe can be privately efficient for Sofia and still be socially costly.

Today: where do the costs show up?

I. Across the economy

Does corruption make countries and firms poorer?

II. In public services

What happens when money meant for roads disappears?

III. Across people

Who bears the burden — and who gets shut out?

We move from Sofia’s private gain to corruption’s wider costs.

I. Costs across the economy

Was Sofia’s $40 grease — or sand?

The “grease” story

  • Without the two payments, delay costs her about $500.
  • With $40 in bribes, the cargo clears and the permit arrives the same day.
  • Sofia is much better off paying than refusing.
  • Corruption appears to help her overcome a slow bureaucracy.

Grease: Leff (1964); Huntington (1968). Sand: Rose-Ackerman (1978); Shleifer and Vishny (1993). Sofia is an illustration, not evidence from these studies.

The fact that paying helps Sofia does not tell us whether the system itself is efficient.

A split view of Sofia at the same permit office: one side shows her moving rapidly through a long queue, while the other suggests officials preserving delay so that faster treatment can be sold.

Two ways to interpret Sofia’s shortcut through the permit office.

Was Sofia’s $40 grease — or sand?

The “sand” story

  • But what if officials can profit from keeping the barrier and queue slow?
  • Then selling speed gives them a reason to preserve the obstacle.
  • Sofia still benefits from paying — while traders who cannot pay remain stuck or stay out.

Grease: Leff (1964); Huntington (1968). Sand: Rose-Ackerman (1978); Shleifer and Vishny (1993). Sofia is an illustration, not evidence from these studies.

The fact that paying helps Sofia does not tell us whether the system itself is efficient.

A split view of Sofia at the same permit office: one side shows her moving rapidly through a long queue, while the other suggests officials preserving delay so that faster treatment can be sold.

Two ways to interpret Sofia’s shortcut through the permit office.

The pattern everyone starts from

Mauro (1995): take Sofia’s question to 67 countries

  • Sofia showed us why a bribe can look like grease for one firm.
  • Mauro asks: does that logic hold across whole economies?
  • He compares 67 countries.
  • Business analysts had rated how much firms faced:
    • corruption,
    • red tape,
    • and an inefficient legal system.

If corruption really greases the wheels, it should be especially useful where red tape is worst.

Mauro (1995), pp. 683–687. Business International collected the institutional ratings in 1980–1983; higher scores meant better institutions.

What would the “grease” story predict?

If bureaucracy is already fast

  • There is little delay to bypass.
  • A bribe should not help very much.

If bureaucracy is painfully slow

  • There is lots of delay to bypass.
  • This is where “speed money” should help most.

So Mauro compares countries with more and less red tape.

Mauro’s world of bureaucratic efficiency

What happens to investment as bureaucracies improve?

Countries with more efficient bureaucracies tended to invest more.

But this picture alone cannot tell us why.

Does corruption help when red tape is worst?

  • Mauro separates countries with more and less red tape.
  • If the grease story is right, corruption should be less harmful — perhaps even helpful — where bureaucracy is slow.
  • It was not.
  • Less corruption was associated with more investment in both groups.

The evidence does not show corruption rescuing countries from bad bureaucracy.

A one-standard-deviation improvement in the corruption score was associated with investment about 2.9 percentage points of GDP higher.

Mauro (1995), Table IV. The estimated relationship between corruption and investment did not differ significantly between high- and low-red-tape countries.

But what causes what?

We observe: Worse bureaucracy ↔︎ less investment

That could mean:

1. Bad bureaucracy → less investment Bribes and red tape make firms less willing to invest.

2. Weak economies → bad bureaucracy Poor economic performance may make effective government harder to sustain.

3. Something else → both Political instability or other institutions may produce both.

The correlation alone cannot tell us which story is right.

Mauro tries to get closer to causality

  • The problem: today’s economy can affect today’s institutions.
  • Mauro therefore looks for something determined much earlier.
  • He uses a country’s ethnolinguistic fractionalization in 1960 to predict institutional quality.
  • Then he asks whether that predicted institutional quality is related to investment.
  • The negative relationship survives.

But the strategy works only if fractionalization affects investment mainly through institutions.

Mauro (1995), pp. 683, 692–698. Mauro also discusses alternative instruments and limitations of the strategy.

The detour, in one picture

%%{init: {"flowchart": {"useMaxWidth": true, "nodeSpacing": 100, "rankSpacing": 95, "padding": 44, "wrappingWidth": 700, "curve": "basis"}, "themeVariables": {"fontSize": "20px", "fontFamily": "Verdana, Geneva, sans-serif"}, "markdownAutoWrap": false, "width": 1150, "height": 620}}%%
flowchart LR
  F["Ethnolinguistic split<br/>in 1960"] -->|"shapes"| I["Institutions today<br/>corruption, red tape"]
  I -->|"related to"| V["Investment today"]
  F -. "assumed: no direct route" .-> V
  classDef early fill:#1e293b,stroke:#1e293b,color:#f9fafb,stroke-width:2px;
  classDef mid fill:#f9fafb,stroke:#4a7c6f,color:#1e293b,stroke-width:2.5px;
  classDef out fill:#f9fafb,stroke:#64748b,color:#1e293b,stroke-width:2.5px;
  class F early;
  class I mid;
  class V out;
  linkStyle 2 stroke:#b44527,stroke-width:2px,stroke-dasharray:6 6;

The 1960 split came first — so that arrow can only run one way.

Author’s illustration of Mauro’s strategy. The solid path is what he can estimate. The dashed arrow is the assumption behind it: fractionalization must not reach investment by any other route. Timing rules out the reverse story; nothing rules out the dashed arrow except argument.

From Mauro’s countries to Ugandan firms

  • Mauro showed that cleaner bureaucracies are associated with more investment across countries.
  • But that evidence was:
    • cross-country,
    • based on perception measures,
    • and still vulnerable to causal doubts.
  • Fisman and Svensson (2007) move to the firm level.
  • They ask: within one country, do firms that pay more bribes grow more slowly?

Same question as Mauro — but now inside one economy, with direct reports from firms.

Fisman and Svensson (2007), Introduction. The paper explicitly presents itself as a firm-level complement to the cross-country literature beginning with Mauro (1995).

Fisman and Svensson (2007): the setup

  • Survey of 243 manufacturing firms in Uganda
  • Managers were asked, confidentially, about:
    • bribe payments,
    • tax payments,
    • and firm performance
  • Bribes and taxes were both measured as a share of sales
  • Outcome: firm growth, especially growth in sales

So the comparison is not just bribes versus no bribes. It is:

How harmful are bribes compared with an ordinary tax burden?

Fisman and Svensson (2007), Abstract and data section. Growth is measured over 1995–1997.

What they found

  • Firms facing higher bribe rates grew more slowly.
  • Firms facing higher tax rates also grew more slowly.
  • But the bribe effect was much larger.
  • Roughly:
    • +1 percentage point in bribery → about 3 percentage points less growth
    • the effect of taxation was much smaller

Bribery looked more damaging to firm growth than taxation.

Fisman and Svensson (2007), Abstract. The paper reports that the bribery effect is about 2.5 to 3 times the size of the tax effect, depending on specification and treatment of outliers.

Why can a bribe hurt more than a tax?

A tax A bribe
Known in advance Secret and uncertain
Applies by rule Depends on the official
Easier to budget for Can be renegotiated or demanded again
Legally enforceable No contract can be enforced in court
May fund public goods Mainly buys a private shortcut

So corruption is not simply “a tax whose revenue disappears.”

It also adds uncertainty, bargaining, and repeat extraction.

Shleifer and Vishny (1993), tested at the firm level by Fisman and Svensson (2007).

When does grease become drag?

Back to Sofia:

Sofia’s first two bribes helped her.

But now imagine her business growing:

  • more shipments;
  • more inspections;
  • more permits and approvals;
  • more points where a payment can be demanded;
  • more uncertainty about how much and when.

A single bribe can help a firm. A system of repeated bribes can slow its growth.

Sofia at a desk with invoices and a calendar while several government counters and possible future demands lie ahead, illustrating the difficulty of planning repeated interactions with officials.

Sofia trying to plan a growing business while facing repeated official demands.

So what?

  • Mauro: across countries, cleaner bureaucracies are associated with more investment.
  • Fisman and Svensson: within one country, firms paying more bribes grow more slowly.
  • And where rents are large, talent can flow into capturing them rather than building things.
  • Together, they suggest that corruption carries real economic costs:
    • across economies,
    • and inside firms.
  • Next we move to a different kind of evidence:
    • corruption in a public project,
    • measured directly,
    • and altered experimentally.

Next: what happens when researchers dig up the road — and some villages face audits by lottery?

Talent diverted into rent-seeking: Murphy, Shleifer and Vishny (1991), cited by Mauro (1995).

II. Costs in public services

A village, a road, and a missing quarter

  • A village in Java receives about US$8,800 to surface its road
  • The village reports, line by line, what it spent
  • Engineers later dig into the road and measure what is actually there
  • About a quarter of the reported spending is not in the ground

Hold that number. First: what part is a transfer, and what part is a social loss?

A two-lane road running through a village in Licin, East Java, with houses and a small truck.

A road in Licin, East Java — one of Olken’s two study provinces.

Photo: Aryphrase, CC BY-SA 3.0, via Wikimedia Commons. Illustrative; not a project road from the study.

Olken (2007). We now use the case to separate money changing hands from public value that disappears.

What exactly was lost?

%%{init: {"flowchart": {"useMaxWidth": true, "nodeSpacing": 105, "rankSpacing": 80, "padding": 48, "wrappingWidth": 700, "curve": "basis"}, "themeVariables": {"fontSize": "20px", "fontFamily": "Verdana, Geneva, sans-serif"}, "markdownAutoWrap": false, "width": 1150, "height": 620}}%%
flowchart LR
  S["$1,000 is<br/>diverted"] --> A["Someone else receives<br/>the $1,000"]
  S --> B["$1,000 less buys<br/>road materials"]
  B --> C["The village gets less road<br/>than it paid for"]
  classDef origin fill:#1e293b,stroke:#1e293b,color:#f9fafb,stroke-width:2px;
  classDef transfer fill:#f9fafb,stroke:#64748b,color:#1e293b,stroke-width:2.5px;
  classDef loss fill:#f9fafb,stroke:#b44527,color:#1e293b,stroke-width:2.5px;
  class S origin;
  class A transfer;
  class B,C loss;

The stolen money changes hands. The social cost is that the public gets less than it paid for.

Author’s illustration. The first branch is a transfer; the second is the loss of public value. Olken (2007) measures missing expenditures and materials, not the downstream effects of a worse road.

Olken (2007): 608 villages in Java

  • Indonesia, 2003–2004
  • A national village-grant programme, funded by the World Bank
  • 608 villages in Central and East Java
  • Most projects: surfacing a dirt road with sand, rocks and gravel

The unit is no longer a country or a firm.

It is a village road.

Where does the comparison come from?

%%{init: {"flowchart": {"useMaxWidth": true, "nodeSpacing": 70, "rankSpacing": 78, "padding": 40, "wrappingWidth": 700, "curve": "basis"}, "themeVariables": {"fontSize": "20px", "fontFamily": "Verdana, Geneva, sans-serif"}, "markdownAutoWrap": false, "width": 1150, "height": 620}}%%
flowchart LR
  A["608 village<br/>road projects"] --> L{"Lottery"}
  L --> T["Told: an audit<br/>is certain"]
  L --> C["Ordinary oversight<br/>4% chance of an audit"]
  T --> M["Engineers measure<br/>what is in the road"]
  C --> M
  M --> D["Difference =<br/>effect of the audit"]
  classDef origin fill:#1e293b,stroke:#1e293b,color:#f9fafb,stroke-width:2px;
  classDef draw fill:#f9fafb,stroke:#b7943a,color:#1e293b,stroke-width:2.5px;
  classDef arm fill:#f9fafb,stroke:#64748b,color:#1e293b,stroke-width:2.5px;
  classDef out fill:#f9fafb,stroke:#b44527,color:#1e293b,stroke-width:2.5px;
  class A origin;
  class L draw;
  class T,C,M arm;
  class D out;

Mauro needed a detour through 1960. Here, the comparison is made by a lottery.

Olken (2007), Sections II–III. Villages were told about the audit before the project was built, so the announcement could change behaviour.

How did they know what was really there?

%%{init: {"flowchart": {"useMaxWidth": true, "nodeSpacing": 45, "rankSpacing": 95, "padding": 44, "wrappingWidth": 700, "curve": "basis"}, "themeVariables": {"fontSize": "20px", "fontFamily": "Verdana, Geneva, sans-serif"}, "markdownAutoWrap": false, "width": 1150, "height": 620}}%%
flowchart LR
  R["The village's own<br/>spending report"] --> G["Compare"]
  S["Core samples dug<br/>out of the road"] --> G
  W["Interviews with<br/>the workers"] --> G
  G --> X["About 24% of reported<br/>spending cannot be found"]
  classDef claim fill:#1e293b,stroke:#1e293b,color:#f9fafb,stroke-width:2px;
  classDef check fill:#f9fafb,stroke:#4a7c6f,color:#1e293b,stroke-width:2.5px;
  classDef mid fill:#f9fafb,stroke:#64748b,color:#1e293b,stroke-width:2.5px;
  classDef hot fill:#b44527,stroke:#b44527,color:#f9fafb,stroke-width:2.5px;
  class R claim;
  class S,W check;
  class G mid;
  class X hot;

Not an expert’s impression of a country. A number you can dig out of the ground.

Olken (2007), pp. 202–203. Engineers took core samples from the finished road, priced materials with local suppliers, and asked workers what they had been paid.

Audits cut the theft

What we learned, and what we did not

What we learned

  • The comparison was made by a lottery, not by choice
  • Announced audits cut missing spending by about a third
  • The theft prevented was worth more than the audits cost

What we did not learn

  • Audits did not end corruption: about a fifth still went missing
  • Village accountability meetings had no average effect
  • One programme, one country, one kind of road

Top-down audits moved the number. Bottom-up village meetings, on average, did not.

Olken (2007), Sections V and VI. Theft prevented ≈ $468 per village against an audit cost of ≈ $335.

III. When corruption becomes exclusion

Sofia could pay. What if you cannot?

  • Sofia faced a $40 shortcut and could afford it.
  • Paying was privately rational: she avoided about $500 in losses.
  • But dealing with the state is not always a choice.
  • And not everyone can buy the shortcut.

The same payment is a shortcut for one person and a barrier for another.

How corruption becomes a personal cost

%%{init: {"flowchart": {"useMaxWidth": true, "nodeSpacing": 45, "rankSpacing": 75, "padding": 40, "wrappingWidth": 700, "curve": "basis"}, "themeVariables": {"fontSize": "20px", "fontFamily": "Verdana, Geneva, sans-serif"}, "markdownAutoWrap": false, "width": 1150, "height": 620}}%%
flowchart LR
  M["Misfortune, or<br/>ordinary business"] --> S["You need<br/>a public office"]
  S --> B["An official asks<br/>for a payment"]
  B --> P["You can pay"]
  B --> N["You cannot pay"]
  P --> A["You get the service,<br/>and pay extra"]
  N --> X["You delay, drop it,<br/>or go without"]
  classDef origin fill:#1e293b,stroke:#1e293b,color:#f9fafb,stroke-width:2px;
  classDef step fill:#f9fafb,stroke:#64748b,color:#1e293b,stroke-width:2.5px;
  classDef pay fill:#f9fafb,stroke:#4a7c6f,color:#1e293b,stroke-width:2.5px;
  classDef gone fill:#f9fafb,stroke:#b44527,color:#1e293b,stroke-width:2.5px;
  class M origin;
  class S,B step;
  class P,A pay;
  class N,X gone;

Three separate questions: who ends up at the counter, what the payment costs them, and what happens if they walk away.

Author’s illustration. We will call these exposure, burden and exit, and take them one at a time.

Exercise · You are in charge

Your city has a corruption problem. Your group controls $1 million.

Groups of four — 5 minutes. Pick one area you care about:

Police · Hospitals · Permits · Construction · Universities · your own

Then answer three questions:

  1. What is the biggest harm corruption causes there?
  2. What one thing would you spend the million on to reduce it?
  3. What from today makes you think that would work?

Report back — 30 seconds each

We chose ______

The main problem is ______

We would spend the money on ______ because ______

Then we vote: which proposal would you actually fund?

No group may vote for itself.

One puzzle, four answers

  • Sofia: a bribe can be privately beneficial.
  • Firms and economies: corruption can coexist with lower investment and slower growth.
  • Public services: diverted resources can leave society with less than it paid for.
  • Households: unofficial payments can turn access to the state into exclusion.

The cost of corruption is not only the money that changes hands. It is what the system makes people stop doing.

Next session: if individually rational payments can sustain socially costly systems, why is corruption so hard to remove?

References

Fisman, R., & Svensson, J. (2007). Are corruption and taxation really harmful to growth? Firm level evidence. Journal of Development Economics, 83(1), 63–75. https://doi.org/10.1016/j.jdeveco.2005.09.009

Hunt, J. (2007). How corruption hits people when they are down. Journal of Development Economics, 84(2), 574–589. https://doi.org/10.1016/j.jdeveco.2007.02.003

Hunt, J., & Laszlo, S. (2012). Is bribery really regressive? Bribery’s costs, benefits, and mechanisms. World Development, 40(2), 355–372. https://doi.org/10.1016/j.worlddev.2011.07.015

Huntington, S. P. (1968). Political order in changing societies. Yale University Press.

Leff, N. H. (1964). Economic development through bureaucratic corruption. American Behavioral Scientist, 8(3), 8–14. https://doi.org/10.1177/000276426400800303

References (continued)

Mauro, P. (1995). Corruption and growth. The Quarterly Journal of Economics, 110(3), 681–712. https://doi.org/10.2307/2946696

Murphy, K. M., Shleifer, A., & Vishny, R. W. (1991). The allocation of talent: Implications for growth. The Quarterly Journal of Economics, 106(2), 503–530.

Olken, B. A. (2007). Monitoring corruption: Evidence from a field experiment in Indonesia. Journal of Political Economy, 115(2), 200–249. https://doi.org/10.1086/517935

Olken, B. A. (2009). Corruption perceptions vs. corruption reality. Journal of Public Economics, 93(7–8), 950–964. https://doi.org/10.1016/j.jpubeco.2009.03.001

Rose-Ackerman, S. (1978). Corruption: A study in political economy. Academic Press.

References (continued)

Shleifer, A., & Vishny, R. W. (1993). Corruption. The Quarterly Journal of Economics, 108(3), 599–617. https://doi.org/10.2307/2118402

Transparency International. (2025). Corruption Perceptions Index 2024. https://www.transparency.org/en/cpi/2024

Treisman, D. (2007). What have we learned about the causes of corruption from ten years of cross-national empirical research? Annual Review of Political Science, 10, 211–244. https://doi.org/10.1146/annurev.polisci.10.081205.095418